{ "schema_version": 1, "repo": "yangzijing/nine-source-hand-data-review-videos-20s", "revision": "8db627faff1bd5bed598abd27f0bd83311193e50", "built_at": "2026-09-20T16:24:49.090469+00:00", "allowed_roots": [ "20_seconds", "short_originals_under_20s" ], "datasets": [ { "dataset": "H2O", "files": 50, "samples": 50, "aliases": 0, "sequences": 50, "twenty_second_files": 21, "short_files": 29, "duration_min_s": 11.7, "duration_median_s": 18.717, "duration_max_s": 20.0, "unique_duration_total_s": 879.307, "modality_metadata": 50, "recording_metadata": 50, "motion_reference": 50, "motion_camera_filter": 50, "motion_hand_filter": 50, "motion_relative_filter": 50 }, { "dataset": "HOT3D", "files": 37, "samples": 37, "aliases": 0, "sequences": 37, "twenty_second_files": 37, "short_files": 0, "duration_min_s": 20.0, "duration_median_s": 20.0, "duration_max_s": 20.0, "unique_duration_total_s": 740.0, "modality_metadata": 37, "recording_metadata": 37, "motion_reference": 37, "motion_camera_filter": 37, "motion_hand_filter": 37, "motion_relative_filter": 37 }, { "dataset": "HOI4D", "files": 50, "samples": 50, "aliases": 0, "sequences": 50, "twenty_second_files": 50, "short_files": 0, "duration_min_s": 20.0, "duration_median_s": 20.0, "duration_max_s": 20.0, "unique_duration_total_s": 1000.0, "modality_metadata": 50, "recording_metadata": 50, "motion_reference": 0, "motion_camera_filter": 0, "motion_hand_filter": 0, "motion_relative_filter": 0 }, { "dataset": "ARCTIC", "files": 94, "samples": 88, "aliases": 6, "sequences": 80, "twenty_second_files": 89, "short_files": 5, "duration_min_s": 19.134, "duration_median_s": 20.0, "duration_max_s": 20.0, "unique_duration_total_s": 1758.002, "modality_metadata": 88, "recording_metadata": 88, "motion_reference": 88, "motion_camera_filter": 88, "motion_hand_filter": 41, "motion_relative_filter": 41 }, { "dataset": "Ego4D", "files": 5, "samples": 5, "aliases": 0, "sequences": 5, "twenty_second_files": 5, "short_files": 0, "duration_min_s": 20.0, "duration_median_s": 20.0, "duration_max_s": 20.0, "unique_duration_total_s": 100.0, "modality_metadata": 5, "recording_metadata": 5, "motion_reference": 0, "motion_camera_filter": 0, "motion_hand_filter": 0, "motion_relative_filter": 0 }, { "dataset": "EgoDex", "files": 46, "samples": 46, "aliases": 0, "sequences": 46, "twenty_second_files": 8, "short_files": 38, "duration_min_s": 2.1, "duration_median_s": 10.0, "duration_max_s": 20.0, "unique_duration_total_s": 452.99999999999994, "modality_metadata": 46, "recording_metadata": 46, "motion_reference": 0, "motion_camera_filter": 0, "motion_hand_filter": 0, "motion_relative_filter": 0 }, { "dataset": "EPIC-KITCHENS", "files": 27, "samples": 27, "aliases": 0, "sequences": 10, "twenty_second_files": 27, "short_files": 0, "duration_min_s": 20.0, "duration_median_s": 20.02, "duration_max_s": 20.02, "unique_duration_total_s": 540.4399999999998, "modality_metadata": 27, "recording_metadata": 27, "motion_reference": 0, "motion_camera_filter": 0, "motion_hand_filter": 0, "motion_relative_filter": 0 }, { "dataset": "HO3D", "files": 50, "samples": 50, "aliases": 0, "sequences": 35, "twenty_second_files": 50, "short_files": 0, "duration_min_s": 20.0, "duration_median_s": 20.0, "duration_max_s": 20.0, "unique_duration_total_s": 1000.0, "modality_metadata": 50, "recording_metadata": 50, "motion_reference": 0, "motion_camera_filter": 0, "motion_hand_filter": 0, "motion_relative_filter": 0 }, { "dataset": "DexYCB", "files": 50, "samples": 50, "aliases": 0, "sequences": 50, "twenty_second_files": 0, "short_files": 50, "duration_min_s": 2.334, "duration_median_s": 2.4, "duration_max_s": 2.467, "unique_duration_total_s": 120.94100000000005, "modality_metadata": 50, "recording_metadata": 50, "motion_reference": 50, "motion_camera_filter": 50, "motion_hand_filter": 26, "motion_relative_filter": 26 } ], "samples": [ { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0bdcd2a298ecb849d96bb379a014f193d4d7a97d496496618067d2263c3aa3b7", "mano_sha256": "77b4b4624aa481a91160dda36922343c5931e0ac743aa9d883231588f1f837e5", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "002647b5b92c791b966af2ed", "sample_id": "124fd3843daad53958ba525d", "file": "20_seconds/H2O/124fd3843daad53958ba525d_MANO_GT_20.00s.mp4", "recording_id": "subject2/o2/1", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "002647b5b92c791b966af2ed", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/124fd3843daad53958ba525d_MANO_GT_20.00s.mp4", "bytes": 6269114, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.027327942863908144, "p90": 0.06984994140067739, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 2.857159594198454, "p90": 8.62642130290429, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0352217462566015, "p90": 0.24380482508471432, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.021726244122742475, "p90": 0.06471139928565853, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.04430305599582123, "p90": 0.21637603688257798, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.04429156327002655, "p90": 0.21634052691880756, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.7350914078941686, "p90": 0.9354698242574039, "valid_pairs": 551, "valid_duration_s": 18.366666666666667, "coverage": 0.9198664440734559 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4883072679661196, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09357586796658256, "p90": 0.2888404025673013, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.023257239190681858, "p90": 0.07876254280874584, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08938230697897045, "p90": 0.26799033701865943, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0893805158979725, "p90": 0.2680625727479783, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.7802155190182335, "p90": 0.94881087292036, "valid_pairs": 591, "valid_duration_s": 19.7, "coverage": 0.986644407345576 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4427542559784701, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06984994140067739, "camera_angular_speed_p90_deg_s": 8.62642130290429, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2888404025673013, "hand_camera_speed_p90_m_s": 0.2680625727479783, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.07876254280874584, "rhcm_p90_m_s": 0.2680625727479783, "rham_p90_deg_s": null, "wms_p90": 0.94881087292036, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "d51797040cb158ac043ee3f21a1ec110ef60fab96fb27dcef22ccf17d5d49c00", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/002647b5b92c791b966af2ed/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0020bac7428581dc962c8f6ab538c9b680cedafdfb505206a666e7fb630d0e80", "mano_sha256": "f36f6193fd5529d9af7ac757a74f4ef8e9cc4fc2301fbc0705828d1d711ca63d", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "03a7f6ce86f30dd7b2d41f5d", "sample_id": "47130401f59f79383516aabf", "file": "20_seconds/H2O/47130401f59f79383516aabf_MANO_GT_20.00s.mp4", "recording_id": "subject3/h1/3", "source_first_frame": "98", "source_last_frame": "697", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "03a7f6ce86f30dd7b2d41f5d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/47130401f59f79383516aabf_MANO_GT_20.00s.mp4", "bytes": 7908335, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.024084277685784427, "p90": 0.04478593775768138, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.2725538717827485, "p90": 7.164453806853557, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.04526647333580331, "p90": 0.1138227729885293, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.020216337023705393, "p90": 0.050385762530160375, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.053529590255092435, "p90": 0.1395293615053427, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.05352877259260736, "p90": 0.13954784803941683, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8246716839237286, "p90": 0.9563471239135871, "valid_pairs": 565, "valid_duration_s": 18.833333333333332, "coverage": 0.9432387312186977 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.22046242064533414, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.17464416254591422, "p90": 0.34705458509512904, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.019094793703215398, "p90": 0.040874585022932224, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.17542559213990463, "p90": 0.34362642448129066, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.17541263040014918, "p90": 0.34361531373357995, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8984627252344095, "p90": 0.978357951067353, "valid_pairs": 584, "valid_duration_s": 19.46666666666667, "coverage": 0.9749582637729549 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.21032758052718856, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04478593775768138, "camera_angular_speed_p90_deg_s": 7.164453806853557, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.34705458509512904, "hand_camera_speed_p90_m_s": 0.34361531373357995, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.050385762530160375, "rhcm_p90_m_s": 0.34361531373357995, "rham_p90_deg_s": null, "wms_p90": 0.978357951067353, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "5d2033216c79e8f5c9fa6578a186600829f7a04035a51062d5613cda0a0a5e08", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/03a7f6ce86f30dd7b2d41f5d/reference.npz", "window_s": [ 3.2666666666666666, 23.233333333333334 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f481380a0a5a719262023fb77426d41106b534bbed9a57777798411edfa03919", "mano_sha256": "0655228ae26c60913dcac54d71dc86d72678de5ac1d5edcf71ff3fb00278cab5", "duration_s": 12.8, "frames": 384, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "0cbccfb55705d602caeea9d1", "sample_id": "422aa345ff5f9e3a354f7a4a", "file": "short_originals_under_20s/H2O/422aa345ff5f9e3a354f7a4a_MANO_GT_12.80s.mp4", "recording_id": "subject2/k1/5", "source_first_frame": "0", "source_last_frame": "383", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "0cbccfb55705d602caeea9d1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/422aa345ff5f9e3a354f7a4a_MANO_GT_12.80s.mp4", "bytes": 4447265, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.019075711245198145, "p90": 0.043307850907635274, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.30759887683909, "p90": 5.2408170740479125, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.05938282864475646, "p90": 0.2528820324456943, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0179459466039463, "p90": 0.0402579251234734, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.061733591712301396, "p90": 0.23567200399689386, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06173132323112431, "p90": 0.23565910274157023, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "wms": { "median": 0.8187821887453017, "p90": 0.9557179861145364, "valid_pairs": 379, "valid_duration_s": 12.633333333333333, "coverage": 0.9895561357702349 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.46093397570844674, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09481356525375662, "p90": 0.2675108134296159, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.018406758859389364, "p90": 0.04053250345535935, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.09615620439823219, "p90": 0.25742963791613804, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.09615511452431186, "p90": 0.257426653181721, "valid_pairs": 383, "valid_duration_s": 12.766666666666667, "coverage": 1.0 }, "wms": { "median": 0.8385582125929556, "p90": 0.9646758456452945, "valid_pairs": 372, "valid_duration_s": 12.400000000000002, "coverage": 0.9712793733681463 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3022231658305196, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.043307850907635274, "camera_angular_speed_p90_deg_s": 5.2408170740479125, "camera_valid_fraction": 1.0, "camera_pairs": 383, "hand_world_speed_p90_m_s": 0.2675108134296159, "hand_camera_speed_p90_m_s": 0.257426653181721, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04053250345535935, "rhcm_p90_m_s": 0.257426653181721, "rham_p90_deg_s": null, "wms_p90": 0.9646758456452945, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "82a6b5ef1a4f9f79110f63ee15992f4276a9f248c81c083f3ef26a52e1ddbe6e", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/0cbccfb55705d602caeea9d1/reference.npz", "window_s": [ 0.0, 12.766666666666667 ], "timestamps_source": "source frame timeline", "frame_count": 384, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "514c24335babf4c818f318e07bc6a2ddf2018070a20325ae7a7ead0ba45e13b7", "mano_sha256": "8ac36aca9a505c507c42bb27d826bf7e2d8ff4f2daf9e381491c704f060fef08", "duration_s": 12.7, "frames": 381, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "13deaba0528deffc474c837e", "sample_id": "b4910660ec2fa33ef5d0b33d", "file": "short_originals_under_20s/H2O/b4910660ec2fa33ef5d0b33d_MANO_GT_12.70s.mp4", "recording_id": "subject2/k1/4", "source_first_frame": "0", "source_last_frame": "380", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "13deaba0528deffc474c837e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/b4910660ec2fa33ef5d0b33d_MANO_GT_12.70s.mp4", "bytes": 4335894, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.023689370840411814, "p90": 0.04991548200394688, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "camera_rotation": { "median": 2.646162235097047, "p90": 5.932252605837407, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0693878438487062, "p90": 0.21670922463713121, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.021172320091830565, "p90": 0.04331122634619028, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06509800674818496, "p90": 0.18698395718236235, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06510827661685134, "p90": 0.18697963326123687, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "wms": { "median": 0.7516402181761919, "p90": 0.9306460269758972, "valid_pairs": 378, "valid_duration_s": 12.6, "coverage": 0.9947368421052631 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3078603474801728, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08162527495434474, "p90": 0.18325048769023816, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01978160625249998, "p90": 0.04271392013032029, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08059873487968658, "p90": 0.18846744653129394, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.08060616149468175, "p90": 0.1884730399793341, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 1.0 }, "wms": { "median": 0.811350475221986, "p90": 0.9468511379173622, "valid_pairs": 379, "valid_duration_s": 12.633333333333333, "coverage": 0.9973684210526316 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.29030045039676244, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04991548200394688, "camera_angular_speed_p90_deg_s": 5.932252605837407, "camera_valid_fraction": 1.0, "camera_pairs": 380, "hand_world_speed_p90_m_s": 0.21670922463713121, "hand_camera_speed_p90_m_s": 0.1884730399793341, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04331122634619028, "rhcm_p90_m_s": 0.1884730399793341, "rham_p90_deg_s": null, "wms_p90": 0.9468511379173622, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "998d0a6e45206da39f7abf63ce74f063fca05eb41753d9407a70e4cd63dfe784", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/13deaba0528deffc474c837e/reference.npz", "window_s": [ 0.0, 12.666666666666666 ], "timestamps_source": "source frame timeline", "frame_count": 381, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b9951acd4a3902790e89a19dd6628c952fa7c1cdeda399f74a7ad1769ff31158", "mano_sha256": "58e35c7c76b34522623d758b823d16f96b5d27c00e747dbedb2355404b893afc", "duration_s": 17.9, "frames": 537, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "1ba57a9bff21a71744e07dab", "sample_id": "fee4e764243b88caff9ca51e", "file": "short_originals_under_20s/H2O/fee4e764243b88caff9ca51e_MANO_GT_17.90s.mp4", "recording_id": "subject3/h1/6", "source_first_frame": "0", "source_last_frame": "536", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "1ba57a9bff21a71744e07dab", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/fee4e764243b88caff9ca51e_MANO_GT_17.90s.mp4", "bytes": 6346959, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.018330993391728986, "p90": 0.03656611268709276, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 1.8362350143153776, "p90": 3.789860245394496, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.033214661529657054, "p90": 0.1185778644974192, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.013123698099264024, "p90": 0.02598916545603449, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03755589589367369, "p90": 0.11287549709082162, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03755041191061392, "p90": 0.11286429149225308, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "wms": { "median": 0.791833926114965, "p90": 0.950841404614322, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 0.9384328358208954 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.08406422438010958, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.0546567400760784, "p90": 0.14532637027475936, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.013405757773476672, "p90": 0.0262759100813893, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05673021019152978, "p90": 0.14830700306055672, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.056731332200400275, "p90": 0.14830283210858247, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 1.0 }, "wms": { "median": 0.8449062780373238, "p90": 0.9615059722251342, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.9925373134328358 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.045661371002243524, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.03656611268709276, "camera_angular_speed_p90_deg_s": 3.789860245394496, "camera_valid_fraction": 1.0, "camera_pairs": 536, "hand_world_speed_p90_m_s": 0.14532637027475936, "hand_camera_speed_p90_m_s": 0.14830283210858247, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.0262759100813893, "rhcm_p90_m_s": 0.14830283210858247, "rham_p90_deg_s": null, "wms_p90": 0.9615059722251342, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "6bff6d3943d7c353bbdc2dd0acc9405cc8e8ea4d810084e65c8cd70ac469920c", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/1ba57a9bff21a71744e07dab/reference.npz", "window_s": [ 0.0, 17.866666666666667 ], "timestamps_source": "source frame timeline", "frame_count": 537, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "fd812986e7d81304b5541c6dfd9331231a46813231c56e2a0865363ee315967a", "mano_sha256": "9ca50dd5e0660dae841239d8aa254a4cbf3da1b99c619a71bb81a66d10707b4f", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "1ca9f4fe734eb500f0ecc373", "sample_id": "a9ffbc7e147da9c3e87211d1", "file": "20_seconds/H2O/a9ffbc7e147da9c3e87211d1_MANO_GT_20.00s.mp4", "recording_id": "subject1/k2/6", "source_first_frame": "77", "source_last_frame": "676", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "1ca9f4fe734eb500f0ecc373", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/a9ffbc7e147da9c3e87211d1_MANO_GT_20.00s.mp4", "bytes": 7044225, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02191871852456749, "p90": 0.06172789737979116, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.3567233958886313, "p90": 7.941267204879774, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.025526063842298486, "p90": 0.0962218239719135, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.014764318532915148, "p90": 0.045290200120141774, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03472825165730362, "p90": 0.08991680669219032, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03472193809090371, "p90": 0.08992055689200704, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.7636352113365861, "p90": 0.9430702034284428, "valid_pairs": 565, "valid_duration_s": 18.833333333333343, "coverage": 0.9432387312186982 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.1388824741347061, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06273903022962411, "p90": 0.24557799947496536, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.015384475228430509, "p90": 0.04217292520392201, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06671122914301841, "p90": 0.23879957280159847, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06672557106714888, "p90": 0.23881447307291387, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8312391133587582, "p90": 0.960790647659086, "valid_pairs": 588, "valid_duration_s": 19.6, "coverage": 0.9816360601001669 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.20990117308111061, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06172789737979116, "camera_angular_speed_p90_deg_s": 7.941267204879774, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.24557799947496536, "hand_camera_speed_p90_m_s": 0.23881447307291387, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.045290200120141774, "rhcm_p90_m_s": 0.23881447307291387, "rham_p90_deg_s": null, "wms_p90": 0.960790647659086, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "bb8dd7233bfe59f456a72a7980330b69a7ad5d28ed83e25d0bc59299166816d5", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/1ca9f4fe734eb500f0ecc373/reference.npz", "window_s": [ 2.566666666666667, 22.533333333333335 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e746152f9623834e6ffe63c592650ceca9d5a57fbeff46f637d8fcf385f306af", "mano_sha256": "041da96883418f41d938c07f07a940897e62ae3a5d82bc37863f0ff09d4c0739", "duration_s": 16.5, "frames": 495, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "2187b4ddb6ea73e48fe9e21d", "sample_id": "fa2a0ac8d54f0cd46c085fe4", "file": "short_originals_under_20s/H2O/fa2a0ac8d54f0cd46c085fe4_MANO_GT_16.50s.mp4", "recording_id": "subject1/k1/5", "source_first_frame": "0", "source_last_frame": "494", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "2187b4ddb6ea73e48fe9e21d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/fa2a0ac8d54f0cd46c085fe4_MANO_GT_16.50s.mp4", "bytes": 5772446, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03317188580398492, "p90": 0.11517425789018243, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 4.515073772129945, "p90": 17.04983236261374, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.038344036555609556, "p90": 0.1479890265873099, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.03503026428115565, "p90": 0.11688905018106462, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06030995242526699, "p90": 0.14373987170498875, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.060363038520681866, "p90": 0.14390110629771743, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "wms": { "median": 0.7791990825978117, "p90": 0.9585309510628074, "valid_pairs": 483, "valid_duration_s": 16.099999999999994, "coverage": 0.9777327935222669 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.1622681720258319, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.12402351434495903, "p90": 0.2822119140535633, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.029684762109559375, "p90": 0.12109396761268527, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.1351749774738048, "p90": 0.31048916455798087, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.13497721322420367, "p90": 0.31048784137216756, "valid_pairs": 494, "valid_duration_s": 16.466666666666665, "coverage": 1.0 }, "wms": { "median": 0.8234040288183069, "p90": 0.9577773585489143, "valid_pairs": 493, "valid_duration_s": 16.43333333333333, "coverage": 0.9979757085020242 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.0568936277060543, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.11517425789018243, "camera_angular_speed_p90_deg_s": 17.04983236261374, "camera_valid_fraction": 1.0, "camera_pairs": 494, "hand_world_speed_p90_m_s": 0.2822119140535633, "hand_camera_speed_p90_m_s": 0.31048784137216756, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.12109396761268527, "rhcm_p90_m_s": 0.31048784137216756, "rham_p90_deg_s": null, "wms_p90": 0.9585309510628074, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "5dcadec8d8b802ecdf03d228b95c61f5ccab1f013b0b690c1631388f2d758989", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/2187b4ddb6ea73e48fe9e21d/reference.npz", "window_s": [ 0.0, 16.466666666666665 ], "timestamps_source": "source frame timeline", "frame_count": 495, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "14c2f6a2662d6024a8ad60d661538626e86aa39561fe62c8d72f2be5eee8ab17", "mano_sha256": "3ba195ebf82c6597e00930c41ffaedd74b74b3827614618047c63abca96acd57", "duration_s": 18.9, "frames": 567, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "23f68a8d9a8cd30cfe5eb15e", "sample_id": "40daedd0aa0a459c89a54109", "file": "short_originals_under_20s/H2O/40daedd0aa0a459c89a54109_MANO_GT_18.90s.mp4", "recording_id": "subject2/h1/4", "source_first_frame": "0", "source_last_frame": "566", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "23f68a8d9a8cd30cfe5eb15e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/40daedd0aa0a459c89a54109_MANO_GT_18.90s.mp4", "bytes": 7357886, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03460862384596892, "p90": 0.0745498650645603, "valid_pairs": 566, "valid_duration_s": 18.866666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.3076816016826527, "p90": 7.2817832307870205, "valid_pairs": 566, "valid_duration_s": 18.866666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08091824428939506, "p90": 0.29203370408202534, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.9081272084805653 }, "cmh_m_s": { "median": 0.021421801169197955, "p90": 0.04860590871923065, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.9081272084805653 }, "rhcm_continuous_estimate_m_s": { "median": 0.08338965019894462, "p90": 0.2790938330228243, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.9081272084805653 }, "rhcm_discrete_m_s": { "median": 0.08342770270509202, "p90": 0.2790773622626771, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.9081272084805653 }, "wms": { "median": 0.8388851720225178, "p90": 0.9610966836337256, "valid_pairs": 512, "valid_duration_s": 17.066666666666663, "coverage": 0.9045936395759715 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.002078894237343703, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.048604863653606126, "p90": 0.13005551756263523, "valid_pairs": 566, "valid_duration_s": 18.866666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.023449138403432504, "p90": 0.06044201672157125, "valid_pairs": 566, "valid_duration_s": 18.866666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05737454651070009, "p90": 0.13988109616413846, "valid_pairs": 566, "valid_duration_s": 18.866666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.057424885780532814, "p90": 0.1400131481378126, "valid_pairs": 566, "valid_duration_s": 18.866666666666667, "coverage": 1.0 }, "wms": { "median": 0.7569705010531257, "p90": 0.9419711501230713, "valid_pairs": 558, "valid_duration_s": 18.6, "coverage": 0.9858657243816255 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.02784965861359745, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0745498650645603, "camera_angular_speed_p90_deg_s": 7.2817832307870205, "camera_valid_fraction": 1.0, "camera_pairs": 566, "hand_world_speed_p90_m_s": 0.29203370408202534, "hand_camera_speed_p90_m_s": 0.2790773622626771, "hand_world_valid_fraction": 0.955026455026455, "hand_camera_valid_fraction": 0.955026455026455, "cmh_p90_m_s": 0.06044201672157125, "rhcm_p90_m_s": 0.2790773622626771, "rham_p90_deg_s": null, "wms_p90": 0.9610966836337256, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "4699836c11ee45ba8cd68b313f46842c28304a99429960ee680d4d67774d3080", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/23f68a8d9a8cd30cfe5eb15e/reference.npz", "window_s": [ 0.0, 18.866666666666667 ], "timestamps_source": "source frame timeline", "frame_count": 567, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e7ae1a683d50bd911a9c56794966847a6248dc715e531c13ae4d0b9ae75f0abc", "mano_sha256": "b1858ea9097ad00d54c0614ba8e90ff3a71b4b7fc9f225ed77466601a06e35a7", "duration_s": 14.9, "frames": 447, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "2512c14df546fcf67caf0914", "sample_id": "1b5dbb0e0973742a56058f4e", "file": "short_originals_under_20s/H2O/1b5dbb0e0973742a56058f4e_MANO_GT_14.90s.mp4", "recording_id": "subject1/o2/0", "source_first_frame": "0", "source_last_frame": "446", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "2512c14df546fcf67caf0914", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/1b5dbb0e0973742a56058f4e_MANO_GT_14.90s.mp4", "bytes": 4723465, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.029631681889076486, "p90": 0.05611012325521571, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.6728197266346143, "p90": 5.667304611592918, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.028817084366341707, "p90": 0.20959547135328116, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.020829399064422112, "p90": 0.04585638980018363, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03520776752955018, "p90": 0.18099810314678344, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.035197019272940146, "p90": 0.1809682782701288, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "wms": { "median": 0.7330594766849934, "p90": 0.93123769319571, "valid_pairs": 401, "valid_duration_s": 13.366666666666669, "coverage": 0.8991031390134531 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.37525100576456877, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1258315208787564, "p90": 0.353599968260526, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.02236005918306032, "p90": 0.04570643007989951, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.11696050757692658, "p90": 0.3470549672876755, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.11697587161228445, "p90": 0.3470540574605535, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "wms": { "median": 0.8349103323513305, "p90": 0.9598515930589798, "valid_pairs": 441, "valid_duration_s": 14.700000000000001, "coverage": 0.9887892376681615 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4064495403257741, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05611012325521571, "camera_angular_speed_p90_deg_s": 5.667304611592918, "camera_valid_fraction": 1.0, "camera_pairs": 446, "hand_world_speed_p90_m_s": 0.353599968260526, "hand_camera_speed_p90_m_s": 0.3470540574605535, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04585638980018363, "rhcm_p90_m_s": 0.3470540574605535, "rham_p90_deg_s": null, "wms_p90": 0.9598515930589798, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "9da629f64b95a92cde3473940d13bdc10c612d1d99f81dc313f8792faa37bd6f", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/2512c14df546fcf67caf0914/reference.npz", "window_s": [ 0.0, 14.866666666666667 ], "timestamps_source": "source frame timeline", "frame_count": 447, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "fc03f0f27fe21c0661d158f18e42727386c79923d9ed46c8f2719ce8c6453e6a", "mano_sha256": "51e897c75fd284e564ba54d3af325d4e58e4fd9d27056b3a0b3d6f75612647c9", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "25e63a664cd15cbb00a95a86", "sample_id": "c4f6626609ae34c3de4d9303", "file": "20_seconds/H2O/c4f6626609ae34c3de4d9303_MANO_GT_20.00s.mp4", "recording_id": "subject1/h2/1", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "25e63a664cd15cbb00a95a86", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/c4f6626609ae34c3de4d9303_MANO_GT_20.00s.mp4", "bytes": 8219298, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03627370557920012, "p90": 0.11783201573844405, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 3.8977184525558637, "p90": 15.797411160765238, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.03553796912970025, "p90": 0.13123907551931965, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.023436904293946723, "p90": 0.09689222892786821, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.04658543620886683, "p90": 0.14697411126799664, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.04656257731426766, "p90": 0.14690590059170155, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.7416697743800426, "p90": 0.9289013953372022, "valid_pairs": 571, "valid_duration_s": 19.03333333333333, "coverage": 0.9532554257095158 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3225919453269384, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08508834500936491, "p90": 0.2974378835145697, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.028255574071885257, "p90": 0.10911076488265778, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.08248955541154648, "p90": 0.273389341406764, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.08254393312074507, "p90": 0.27343423714890686, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.7410790033320912, "p90": 0.9561473243953921, "valid_pairs": 597, "valid_duration_s": 19.9, "coverage": 0.996661101836394 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3426604541571515, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.11783201573844405, "camera_angular_speed_p90_deg_s": 15.797411160765238, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2974378835145697, "hand_camera_speed_p90_m_s": 0.27343423714890686, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.10911076488265778, "rhcm_p90_m_s": 0.27343423714890686, "rham_p90_deg_s": null, "wms_p90": 0.9561473243953921, "gt_coverage": 1.0000000000000002, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "a3e9beb2628952e01c17afae8c5aab80c6b8aacd7a4ce3a8b4715af6d001079b", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/25e63a664cd15cbb00a95a86/reference.npz", "window_s": [ 0.03333333333333333, 20.0 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e8e5271bd9aa571a821da4dc071661bae102547a9f0dfdd3cb5aa0f4a87657be", "mano_sha256": "cc765b2c4e2ed159366f0e65a95afcb088194f22718871adcd3386198b817b4f", "duration_s": 19.567, "frames": 587, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "3ef9fa42b8b8eb9cb652c0a2", "sample_id": "455932daee971aae3e0ff211", "file": "short_originals_under_20s/H2O/455932daee971aae3e0ff211_MANO_GT_19.57s.mp4", "recording_id": "subject1/h1/4", "source_first_frame": "0", "source_last_frame": "586", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "3ef9fa42b8b8eb9cb652c0a2", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/455932daee971aae3e0ff211_MANO_GT_19.57s.mp4", "bytes": 8444634, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03156083928815101, "p90": 0.06383205735507871, "valid_pairs": 586, "valid_duration_s": 19.533333333333335, "coverage": 1.0 }, "camera_rotation": { "median": 3.792259620679272, "p90": 7.441436795567736, "valid_pairs": 586, "valid_duration_s": 19.533333333333335, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03080020014923021, "p90": 0.13288142103048226, "valid_pairs": 586, "valid_duration_s": 19.533333333333335, "coverage": 1.0 }, "cmh_m_s": { "median": 0.023097073751873115, "p90": 0.058406979167892926, "valid_pairs": 586, "valid_duration_s": 19.533333333333335, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.041633586921091946, "p90": 0.13456196990510338, "valid_pairs": 586, "valid_duration_s": 19.533333333333335, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.041624531884352126, "p90": 0.1345324991561722, "valid_pairs": 586, "valid_duration_s": 19.533333333333335, "coverage": 1.0 }, "wms": { "median": 0.7221487377649041, "p90": 0.9347247907507443, "valid_pairs": 570, "valid_duration_s": 18.999999999999993, "coverage": 0.972696245733788 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.2466053605369034, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06807102053726875, "p90": 0.36433944575111993, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9470989761092149 }, "cmh_m_s": { "median": 0.025144222304067383, "p90": 0.04874647342416882, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9470989761092149 }, "rhcm_continuous_estimate_m_s": { "median": 0.0732270870718016, "p90": 0.3615462132269362, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9470989761092149 }, "rhcm_discrete_m_s": { "median": 0.07323851343255214, "p90": 0.3616986845457861, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9470989761092149 }, "wms": { "median": 0.8327626739454742, "p90": 0.9545820716277388, "valid_pairs": 547, "valid_duration_s": 18.233333333333334, "coverage": 0.9334470989761092 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.35128361356187754, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06383205735507871, "camera_angular_speed_p90_deg_s": 7.441436795567736, "camera_valid_fraction": 1.0, "camera_pairs": 586, "hand_world_speed_p90_m_s": 0.36433944575111993, "hand_camera_speed_p90_m_s": 0.3616986845457861, "hand_world_valid_fraction": 0.9744463373083475, "hand_camera_valid_fraction": 0.9744463373083475, "cmh_p90_m_s": 0.058406979167892926, "rhcm_p90_m_s": 0.3616986845457861, "rham_p90_deg_s": null, "wms_p90": 0.9545820716277388, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "7e4ec6b3e59930e3eed16a4747b443df18645e98e0d1890d29f9d9182711e94a", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/3ef9fa42b8b8eb9cb652c0a2/reference.npz", "window_s": [ 0.0, 19.533333333333335 ], "timestamps_source": "source frame timeline", "frame_count": 587, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c9637579b78b6837f2dba2d0e11686a0594bdb304416c689a80abfc296c15dbf", "mano_sha256": "49322951c07d3a0c1a6ec10ead82e031b8a0b0d1f3381deae50f29b8594ec9a4", "duration_s": 13.034, "frames": 391, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "4491ae0fc4247501e4f29095", "sample_id": "83b3cce336095e816b298b80", "file": "short_originals_under_20s/H2O/83b3cce336095e816b298b80_MANO_GT_13.03s.mp4", "recording_id": "subject2/o2/2", "source_first_frame": "0", "source_last_frame": "390", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "4491ae0fc4247501e4f29095", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/83b3cce336095e816b298b80_MANO_GT_13.03s.mp4", "bytes": 4099447, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.024796989152820054, "p90": 0.046838455294486235, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "camera_rotation": { "median": 3.1468338251601664, "p90": 6.943542406552774, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06302465526524005, "p90": 0.2392565298642329, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "cmh_m_s": { "median": 0.022212386093662245, "p90": 0.052750876650444214, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06544968475152842, "p90": 0.235363570569819, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06544444795018714, "p90": 0.2353042167469769, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "wms": { "median": 0.7795577922107451, "p90": 0.951697934481661, "valid_pairs": 375, "valid_duration_s": 12.5, "coverage": 0.9615384615384616 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3650833692832654, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07344123131202104, "p90": 0.22876657124387956, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "cmh_m_s": { "median": 0.020962066298450023, "p90": 0.049586133783842516, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08609649268479778, "p90": 0.22494061109694133, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.08610530419980542, "p90": 0.2249487479851949, "valid_pairs": 390, "valid_duration_s": 13.0, "coverage": 1.0 }, "wms": { "median": 0.8351920354281441, "p90": 0.9628504665370476, "valid_pairs": 384, "valid_duration_s": 12.799999999999999, "coverage": 0.9846153846153846 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.32706683450606816, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.046838455294486235, "camera_angular_speed_p90_deg_s": 6.943542406552774, "camera_valid_fraction": 1.0, "camera_pairs": 390, "hand_world_speed_p90_m_s": 0.2392565298642329, "hand_camera_speed_p90_m_s": 0.2353042167469769, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.052750876650444214, "rhcm_p90_m_s": 0.2353042167469769, "rham_p90_deg_s": null, "wms_p90": 0.9628504665370476, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "af799b6639485882cf9826259195e239cf084e81feccae8997b40dff5e32336c", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/4491ae0fc4247501e4f29095/reference.npz", "window_s": [ 0.0, 13.0 ], "timestamps_source": "source frame timeline", "frame_count": 391, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "bbc5f81227d272c9415e06643efd86dc05394679deb5349d68ad5745c2812ad7", "mano_sha256": "2bd9228161b2a00978306302957894eb0f43104875b360d9b2879801afb447f9", "duration_s": 18.1, "frames": 543, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "48b5c70e03db9ce5d8b09656", "sample_id": "54f1dbd77328b7debad47b80", "file": "short_originals_under_20s/H2O/54f1dbd77328b7debad47b80_MANO_GT_18.10s.mp4", "recording_id": "subject1/o1/4", "source_first_frame": "0", "source_last_frame": "542", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "48b5c70e03db9ce5d8b09656", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/54f1dbd77328b7debad47b80_MANO_GT_18.10s.mp4", "bytes": 6513202, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.038048048729872365, "p90": 0.07810653824372517, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 1.0 }, "camera_rotation": { "median": 3.464519057731031, "p90": 7.232421346533975, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.023543216364360353, "p90": 0.09088638105395572, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.023745156392160446, "p90": 0.05981962014024718, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.039116601248877045, "p90": 0.09269313943963808, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.039148920350025804, "p90": 0.09270120370615532, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 1.0 }, "wms": { "median": 0.757642223015492, "p90": 0.947126128974042, "valid_pairs": 513, "valid_duration_s": 17.1, "coverage": 0.9464944649446495 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.12525275006122677, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.0686258646784415, "p90": 0.45757318110289247, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.9354243542435424 }, "cmh_m_s": { "median": 0.025300305747213772, "p90": 0.0651499695199945, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.9354243542435424 }, "rhcm_continuous_estimate_m_s": { "median": 0.07782728294623774, "p90": 0.4565250824478323, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.9354243542435424 }, "rhcm_discrete_m_s": { "median": 0.07781862398617535, "p90": 0.45653133975065746, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.9354243542435424 }, "wms": { "median": 0.8115211825020427, "p90": 0.965040883488936, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.92619926199262 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.19015562953360865, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.07810653824372517, "camera_angular_speed_p90_deg_s": 7.232421346533975, "camera_valid_fraction": 1.0, "camera_pairs": 542, "hand_world_speed_p90_m_s": 0.45757318110289247, "hand_camera_speed_p90_m_s": 0.45653133975065746, "hand_world_valid_fraction": 0.9686924493554327, "hand_camera_valid_fraction": 0.9686924493554327, "cmh_p90_m_s": 0.0651499695199945, "rhcm_p90_m_s": 0.45653133975065746, "rham_p90_deg_s": null, "wms_p90": 0.965040883488936, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "32ef7210ea92806f3c0f44f107f0605f02b636bd98c5bc7b45956a9ed124212a", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/48b5c70e03db9ce5d8b09656/reference.npz", "window_s": [ 0.0, 18.066666666666666 ], "timestamps_source": "source frame timeline", "frame_count": 543, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8b0e309e16542b5cd311e8b1bea53d6ce9ea685110bda266c35fe87c4b882e37", "mano_sha256": "5553816815496258155660f1e0bb7fe59811865b0f2992a9b19e5e2ce3a4a7dc", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "55779f727fd4970e0e421510", "sample_id": "7ac4ca051928b11e33b71c8f", "file": "20_seconds/H2O/7ac4ca051928b11e33b71c8f_MANO_GT_20.00s.mp4", "recording_id": "subject2/h1/5", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "55779f727fd4970e0e421510", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/7ac4ca051928b11e33b71c8f_MANO_GT_20.00s.mp4", "bytes": 8113057, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.032759468397357244, "p90": 0.06073261272694186, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 3.28759109666251, "p90": 6.364038230800262, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07402498018842855, "p90": 0.19650991575562224, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.956594323873122 }, "cmh_m_s": { "median": 0.02441775673083922, "p90": 0.05712891944097618, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.956594323873122 }, "rhcm_continuous_estimate_m_s": { "median": 0.07655153636173292, "p90": 0.205071785895662, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.956594323873122 }, "rhcm_discrete_m_s": { "median": 0.07655912832547455, "p90": 0.20507560787255158, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.956594323873122 }, "wms": { "median": 0.7990853458577103, "p90": 0.9564898208813752, "valid_pairs": 561, "valid_duration_s": 18.700000000000003, "coverage": 0.9365609348914861 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.2996456997223183, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.15269045566905415, "p90": 0.3402372543710782, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.024885789030733563, "p90": 0.05184733461809606, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.14488742558891932, "p90": 0.35001200896648815, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.1448767165576984, "p90": 0.3500249654628859, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.8651638060374583, "p90": 0.9653189736566314, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3642688203862624, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06073261272694186, "camera_angular_speed_p90_deg_s": 6.364038230800262, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3402372543710782, "hand_camera_speed_p90_m_s": 0.3500249654628859, "hand_world_valid_fraction": 0.9783333333333334, "hand_camera_valid_fraction": 0.9783333333333334, "cmh_p90_m_s": 0.05712891944097618, "rhcm_p90_m_s": 0.3500249654628859, "rham_p90_deg_s": null, "wms_p90": 0.9653189736566314, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "71901886c907343fa4c7b41f5bf21c8a8a0ca4d30b1ba96f32f17a31371ca098", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/55779f727fd4970e0e421510/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "adad899d44121a03237ef5b166e732d4009914a4b4c2ce6eac46b91b87cb8f70", "mano_sha256": "b92eab04482e5ed8e7993882bcb6b5c9b2e90f081c9f11cf32512235949c0183", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "579b02e6b1d47fa7db1d9fd9", "sample_id": "1c3d4d3c4fdc4cffaeaf4c4f", "file": "20_seconds/H2O/1c3d4d3c4fdc4cffaeaf4c4f_MANO_GT_20.00s.mp4", "recording_id": "subject1/o2/2", "source_first_frame": "405", "source_last_frame": "1004", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "579b02e6b1d47fa7db1d9fd9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/1c3d4d3c4fdc4cffaeaf4c4f_MANO_GT_20.00s.mp4", "bytes": 5766780, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0200014541911332, "p90": 0.042362897627521344, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 1.9644693983278612, "p90": 3.5117458634945806, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03239735012504309, "p90": 0.11129921802682256, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.013006814348566095, "p90": 0.02792837879225796, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03466949183301528, "p90": 0.11383943173839722, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03467142619128537, "p90": 0.1138385517594573, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.7989226598173024, "p90": 0.9566425359417647, "valid_pairs": 567, "valid_duration_s": 18.900000000000013, "coverage": 0.9465776293823044 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.0043556429053997455, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07711586425070781, "p90": 0.20073448018630313, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.012789314222864489, "p90": 0.029135201944952966, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.07288129951986949, "p90": 0.19355586841436606, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0728816805964276, "p90": 0.19354635009429233, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8581094266474728, "p90": 0.9761346043303677, "valid_pairs": 596, "valid_duration_s": 19.866666666666667, "coverage": 0.9949916527545909 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.06502141249909267, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.042362897627521344, "camera_angular_speed_p90_deg_s": 3.5117458634945806, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.20073448018630313, "hand_camera_speed_p90_m_s": 0.19354635009429233, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.029135201944952966, "rhcm_p90_m_s": 0.19354635009429233, "rham_p90_deg_s": null, "wms_p90": 0.9761346043303677, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "f5f3cc295ffbfc2457cf97df3a7942996c6e6ec7a12d6a2215874c92e0b9f66f", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/579b02e6b1d47fa7db1d9fd9/reference.npz", "window_s": [ 13.5, 33.46666666666667 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "41fb4a13312fa3edd2cf4ec8aaeb1f65f41dee239fca6b370e24f306a735475c", "mano_sha256": "cb8b523d4c19101c27f36fd7e8983338ae182499d41b39be75caf298c7d789c1", "duration_s": 15.967, "frames": 479, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "5b62f99fd1df9a4b04e85e7c", "sample_id": "930adbaa8ed9c44ae4ceccb3", "file": "short_originals_under_20s/H2O/930adbaa8ed9c44ae4ceccb3_MANO_GT_15.97s.mp4", "recording_id": "subject1/h2/5", "source_first_frame": "0", "source_last_frame": "478", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "5b62f99fd1df9a4b04e85e7c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/930adbaa8ed9c44ae4ceccb3_MANO_GT_15.97s.mp4", "bytes": 6426949, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.029797607869685962, "p90": 0.06670502241716396, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "camera_rotation": { "median": 3.8634364328330353, "p90": 10.810547035538441, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.05084955540482115, "p90": 0.1562276576013292, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "cmh_m_s": { "median": 0.029346106953972755, "p90": 0.08176043570505302, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06129567501042741, "p90": 0.16079680719300585, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06123997683579128, "p90": 0.16090864937683164, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "wms": { "median": 0.785029610015092, "p90": 0.9569083519258996, "valid_pairs": 470, "valid_duration_s": 15.666666666666668, "coverage": 0.9832635983263599 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.12181401225624265, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.179909821686637, "p90": 0.3506219368785547, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "cmh_m_s": { "median": 0.027672864309483958, "p90": 0.08437155722860368, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.1780494759351948, "p90": 0.3447272845173596, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.17805976260676915, "p90": 0.34472847149950414, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "wms": { "median": 0.8672699104554649, "p90": 0.9731268009301625, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": -0.006569270042833176, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06670502241716396, "camera_angular_speed_p90_deg_s": 10.810547035538441, "camera_valid_fraction": 1.0, "camera_pairs": 478, "hand_world_speed_p90_m_s": 0.3506219368785547, "hand_camera_speed_p90_m_s": 0.34472847149950414, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.08437155722860368, "rhcm_p90_m_s": 0.34472847149950414, "rham_p90_deg_s": null, "wms_p90": 0.9731268009301625, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "42ccfe56fe638c4828b9bcefb899a9f9b784bf805b5560f00872dc0e545d3937", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/5b62f99fd1df9a4b04e85e7c/reference.npz", "window_s": [ 0.0, 15.933333333333334 ], "timestamps_source": "source frame timeline", "frame_count": 479, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "95b283f01b5a9f691dfb86ed9e24af769798351c23e7dd43617bafbaac92e43e", "mano_sha256": "73c4266868066b3d9f6e8722d6fd82cb23b1ebfe6fdc2bedd40005e825913670", "duration_s": 17.367, "frames": 521, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "5d7f89c6170ce811dfb2807c", "sample_id": "f86f81e33ffcbd42cc016158", "file": "short_originals_under_20s/H2O/f86f81e33ffcbd42cc016158_MANO_GT_17.37s.mp4", "recording_id": "subject2/o1/5", "source_first_frame": "0", "source_last_frame": "520", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "5d7f89c6170ce811dfb2807c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/f86f81e33ffcbd42cc016158_MANO_GT_17.37s.mp4", "bytes": 5113735, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.01982151773454309, "p90": 0.038640836267640234, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "camera_rotation": { "median": 2.1150500767922464, "p90": 4.3627599259907095, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0179972880898063, "p90": 0.06061697473770191, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "cmh_m_s": { "median": 0.014939328211430847, "p90": 0.032996478968794575, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.0263513763446643, "p90": 0.06915442250481914, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.026353903056000872, "p90": 0.06916050533050622, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "wms": { "median": 0.7601222980753802, "p90": 0.9323241201520428, "valid_pairs": 447, "valid_duration_s": 14.900000000000004, "coverage": 0.8596153846153849 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.15367435644248795, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.061200982897547655, "p90": 0.2113688765681821, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01553435081998366, "p90": 0.033836638193623245, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06417196256990697, "p90": 0.21508685502301464, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06416618447537478, "p90": 0.21507109761357315, "valid_pairs": 520, "valid_duration_s": 17.333333333333332, "coverage": 1.0 }, "wms": { "median": 0.8353840540898622, "p90": 0.9589781118034566, "valid_pairs": 513, "valid_duration_s": 17.099999999999998, "coverage": 0.9865384615384615 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.22440103471024436, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.038640836267640234, "camera_angular_speed_p90_deg_s": 4.3627599259907095, "camera_valid_fraction": 1.0, "camera_pairs": 520, "hand_world_speed_p90_m_s": 0.2113688765681821, "hand_camera_speed_p90_m_s": 0.21507109761357315, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.033836638193623245, "rhcm_p90_m_s": 0.21507109761357315, "rham_p90_deg_s": null, "wms_p90": 0.9589781118034566, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "7ed31a9154fbe0e90675b30e954f2d9f94daafccc5be2aac03acf8ca0d18f0cb", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/5d7f89c6170ce811dfb2807c/reference.npz", "window_s": [ 0.0, 17.333333333333332 ], "timestamps_source": "source frame timeline", "frame_count": 521, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ecc62f830da2bce3547d0c8510d8e275415296323329f3516374895ec4dcb74b", "mano_sha256": "367ffe48dba60fa8e14f7adf420998f0ffcd7012b973cfb904dc4730b94c3ff7", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "63e8a16043268c732fd7bb50", "sample_id": "9c1eb25f9f2c65c26c44ffbc", "file": "20_seconds/H2O/9c1eb25f9f2c65c26c44ffbc_MANO_GT_20.00s.mp4", "recording_id": "subject1/k1/3", "source_first_frame": "122", "source_last_frame": "721", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "63e8a16043268c732fd7bb50", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/9c1eb25f9f2c65c26c44ffbc_MANO_GT_20.00s.mp4", "bytes": 6712673, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.030253526807258468, "p90": 0.062415544654800106, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.684700232413698, "p90": 9.518163634993797, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.039518025767536116, "p90": 0.12037368067759706, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.026522826341847185, "p90": 0.06412475599317419, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.054899901910182555, "p90": 0.13440198971449352, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.05496131210563558, "p90": 0.13435993845106525, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.7737401499870009, "p90": 0.951555068584979, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.1806579527752497, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.11793477038690996, "p90": 0.31869842852299335, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.023941510200020902, "p90": 0.051218188855263316, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.11512652780031532, "p90": 0.30356187034239834, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.11523674431928395, "p90": 0.3035620728748416, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8196772205431797, "p90": 0.9549603734853551, "valid_pairs": 594, "valid_duration_s": 19.8, "coverage": 0.9916527545909849 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.39895863228010997, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.062415544654800106, "camera_angular_speed_p90_deg_s": 9.518163634993797, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.31869842852299335, "hand_camera_speed_p90_m_s": 0.3035620728748416, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.06412475599317419, "rhcm_p90_m_s": 0.3035620728748416, "rham_p90_deg_s": null, "wms_p90": 0.9549603734853551, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "b2671dec02efcdb3002d9622a8b225d8b2d987bf2957e4d9a786e077c3dc8935", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/63e8a16043268c732fd7bb50/reference.npz", "window_s": [ 4.066666666666666, 24.033333333333335 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "dd569082c8d37b4eeaa74b842f643c323cfe22430dd9e3d655063fc22e7a8deb", "mano_sha256": "13d1411d1000404a23f2ac95329f3362ff3836a805c7dfc0d388cca56cc17ab5", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "70ad14851b39cfd45a21a87e", "sample_id": "29d0e05feae11d122ef02a40", "file": "20_seconds/H2O/29d0e05feae11d122ef02a40_MANO_GT_20.00s.mp4", "recording_id": "subject2/h2/4", "source_first_frame": "16", "source_last_frame": "615", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "70ad14851b39cfd45a21a87e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/29d0e05feae11d122ef02a40_MANO_GT_20.00s.mp4", "bytes": 7156075, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.020896775307035455, "p90": 0.0410723448471632, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 2.0550827310358817, "p90": 4.342426034476384, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.03409961414799442, "p90": 0.1083452068468795, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.013758161202886813, "p90": 0.028915818716362493, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.03688460961081453, "p90": 0.11156728806815484, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.036875859816269424, "p90": 0.1115663154618665, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.7979487657968941, "p90": 0.9518583037340417, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.24929821720705073, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07117268371902309, "p90": 0.15048108310073158, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.014210139214800558, "p90": 0.029622444208174684, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.06844435135979783, "p90": 0.15368862918381657, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.06844400935581677, "p90": 0.15369068021182075, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.8345916482552598, "p90": 0.9596772912965825, "valid_pairs": 582, "valid_duration_s": 19.4, "coverage": 0.9716193656093489 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.1926669309495759, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0410723448471632, "camera_angular_speed_p90_deg_s": 4.342426034476384, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.15048108310073158, "hand_camera_speed_p90_m_s": 0.15369068021182075, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.029622444208174684, "rhcm_p90_m_s": 0.15369068021182075, "rham_p90_deg_s": null, "wms_p90": 0.9596772912965825, "gt_coverage": 1.0000000000000002, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "4f62850113dc911a189990a10b3621e2303005724b1dde47fd1a5278acc1ff4b", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/70ad14851b39cfd45a21a87e/reference.npz", "window_s": [ 0.5333333333333333, 20.5 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4494d126bb445c412783f04de7d331d4775a72ce1476a20766d1d6ba5917949f", "mano_sha256": "1eddc16836f77645564a54ff9454be474b7ee0e67988eedd5459f6ed470503cd", "duration_s": 16.8, "frames": 504, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "74d554f697930874c3633a73", "sample_id": "6dbaf63ab83deace7519de22", "file": "short_originals_under_20s/H2O/6dbaf63ab83deace7519de22_MANO_GT_16.80s.mp4", "recording_id": "subject3/k1/7", "source_first_frame": "0", "source_last_frame": "503", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "74d554f697930874c3633a73", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/6dbaf63ab83deace7519de22_MANO_GT_16.80s.mp4", "bytes": 7338289, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02201555671390192, "p90": 0.04082221646421188, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "camera_rotation": { "median": 2.8597872780608973, "p90": 5.493850032715327, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.042624144272130286, "p90": 0.2024926069683822, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01726717733951918, "p90": 0.036780483291512865, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.042750823235292855, "p90": 0.1935185280345055, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.04275577790387088, "p90": 0.1935658128382471, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "wms": { "median": 0.7384676478638084, "p90": 0.9236532386808178, "valid_pairs": 474, "valid_duration_s": 15.799999999999994, "coverage": 0.94234592445328 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.41962557796835004, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06889080356977122, "p90": 0.2211376438373902, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.016812388008844832, "p90": 0.03237477490148463, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.07047324699280766, "p90": 0.21037090517106594, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0704735866159387, "p90": 0.21036079117406897, "valid_pairs": 503, "valid_duration_s": 16.766666666666666, "coverage": 1.0 }, "wms": { "median": 0.8222806268669828, "p90": 0.9605068170540406, "valid_pairs": 497, "valid_duration_s": 16.566666666666663, "coverage": 0.9880715705765406 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.20859664626545416, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04082221646421188, "camera_angular_speed_p90_deg_s": 5.493850032715327, "camera_valid_fraction": 1.0, "camera_pairs": 503, "hand_world_speed_p90_m_s": 0.2211376438373902, "hand_camera_speed_p90_m_s": 0.21036079117406897, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.036780483291512865, "rhcm_p90_m_s": 0.21036079117406897, "rham_p90_deg_s": null, "wms_p90": 0.9605068170540406, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "0c2e361e3320f73716a94c2c427bb356a9ca262b7b35dd7510d7a35c56200c19", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/74d554f697930874c3633a73/reference.npz", "window_s": [ 0.0, 16.766666666666666 ], "timestamps_source": "source frame timeline", "frame_count": 504, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8101b5490470f1d7871b8d1bf55ebb369006f772e4c95bb27c60621869e4018a", "mano_sha256": "ef22beac048860ff486f63d9da0074b2bfafb2cd1a4c7b7ee2e0fb231489ae09", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "751706d57d496dca0c2d24b8", "sample_id": "cb736b0be44ef630433285c1", "file": "20_seconds/H2O/cb736b0be44ef630433285c1_MANO_GT_20.00s.mp4", "recording_id": "subject1/k1/1", "source_first_frame": "154", "source_last_frame": "753", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "751706d57d496dca0c2d24b8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/cb736b0be44ef630433285c1_MANO_GT_20.00s.mp4", "bytes": 6480781, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02663694936141869, "p90": 0.08947389914457245, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.769416900003929, "p90": 12.307446318634433, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03165835643066782, "p90": 0.17795426017338692, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.019828387369039607, "p90": 0.07170479035298138, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05486550701462773, "p90": 0.17496358854632926, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.054866168211874934, "p90": 0.17493775565715816, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8209745264247047, "p90": 0.9549772647079645, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098495 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.19920424788247967, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.056253079498876424, "p90": 0.25990016764002316, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.02230676972130918, "p90": 0.09886014385898174, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05981692903635275, "p90": 0.16186796073594986, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.059816285943324964, "p90": 0.16190759644178243, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.7376267005002792, "p90": 0.933277494637775, "valid_pairs": 591, "valid_duration_s": 19.7, "coverage": 0.9866444073455758 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.49891268055454735, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.08947389914457245, "camera_angular_speed_p90_deg_s": 12.307446318634433, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.25990016764002316, "hand_camera_speed_p90_m_s": 0.17493775565715816, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.09886014385898174, "rhcm_p90_m_s": 0.17493775565715816, "rham_p90_deg_s": null, "wms_p90": 0.9549772647079645, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "d586497c863f50d2ff78013919633624546b4ed459a3324019d69803a468e715", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/751706d57d496dca0c2d24b8/reference.npz", "window_s": [ 5.133333333333334, 25.1 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "07423005bfd3baddcf40941570ceed2f916691075f81bfda4be0de8a0374d7d1", "mano_sha256": "1d2a8d0c32abb96fcd11e85979e8dca23272c4bba13f4a2dc4c7e8565aee0104", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "77d1fc3a3d81db44b0a46927", "sample_id": "f6cd1056b8f073581d13b173", "file": "20_seconds/H2O/f6cd1056b8f073581d13b173_MANO_GT_20.00s.mp4", "recording_id": "subject1/k1/2", "source_first_frame": "145", "source_last_frame": "744", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "77d1fc3a3d81db44b0a46927", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/f6cd1056b8f073581d13b173_MANO_GT_20.00s.mp4", "bytes": 5757566, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.01612952268031001, "p90": 0.03494809983226508, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 1.5106348516675214, "p90": 4.456270769178496, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.032661115591342896, "p90": 0.09200124090548426, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.010762687711113755, "p90": 0.029509802965808542, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.0349668782116197, "p90": 0.09856979688655458, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03496562375265465, "p90": 0.09857689464863965, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8056961778501663, "p90": 0.9454795381009554, "valid_pairs": 540, "valid_duration_s": 18.000000000000007, "coverage": 0.901502504173623 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.275480762251467, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.05546160481303604, "p90": 0.19477567832249762, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.010809329064410187, "p90": 0.03142435020970267, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05765206396691204, "p90": 0.18085213998435593, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.05765494596006395, "p90": 0.18084979595110026, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8475280599862898, "p90": 0.9726112202033288, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 0.9899833055091818 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.16342605568924795, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.03494809983226508, "camera_angular_speed_p90_deg_s": 4.456270769178496, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.19477567832249762, "hand_camera_speed_p90_m_s": 0.18084979595110026, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.03142435020970267, "rhcm_p90_m_s": 0.18084979595110026, "rham_p90_deg_s": null, "wms_p90": 0.9726112202033288, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "03a7562d9571a287f18039d28001c7455013be7ecf459afab6eca9e1ead465ee", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/77d1fc3a3d81db44b0a46927/reference.npz", "window_s": [ 4.833333333333333, 24.8 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7228dbaa9765c9f659d404a70bd83312750fa9fb5d5123ea17adc6f47ad61b0f", "mano_sha256": "aefd5ee6d41e381e3d54c00fcf055f4adad59ab88e1ef64dae318c04f079db04", "duration_s": 13.734, "frames": 412, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "7a92636dd7fa26f9c9cacade", "sample_id": "203e5bd4cce0f89cda399029", "file": "short_originals_under_20s/H2O/203e5bd4cce0f89cda399029_MANO_GT_13.73s.mp4", "recording_id": "subject3/h1/2", "source_first_frame": "0", "source_last_frame": "411", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "7a92636dd7fa26f9c9cacade", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/203e5bd4cce0f89cda399029_MANO_GT_13.73s.mp4", "bytes": 4636693, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.019268942660973624, "p90": 0.036213281652152735, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "camera_rotation": { "median": 1.9382834731332383, "p90": 4.130891866219533, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0420579583768649, "p90": 0.17704498140867822, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "cmh_m_s": { "median": 0.014411142557757396, "p90": 0.03247533127102501, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.046284934645678844, "p90": 0.1811133831073384, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0462744767650606, "p90": 0.18111315473770453, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "wms": { "median": 0.8232670145166548, "p90": 0.9512763168983561, "valid_pairs": 376, "valid_duration_s": 12.533333333333333, "coverage": 0.9148418491484185 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3067385589170185, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.05367400163659295, "p90": 0.2090975387407451, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "cmh_m_s": { "median": 0.014869998084915742, "p90": 0.03261977593364048, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05451553249898315, "p90": 0.20574844347366494, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0545463920046391, "p90": 0.20574439734742442, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 1.0 }, "wms": { "median": 0.8238740952128042, "p90": 0.9570865764220556, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 0.9708029197080293 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3166524645191176, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.036213281652152735, "camera_angular_speed_p90_deg_s": 4.130891866219533, "camera_valid_fraction": 1.0, "camera_pairs": 411, "hand_world_speed_p90_m_s": 0.2090975387407451, "hand_camera_speed_p90_m_s": 0.20574439734742442, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.03261977593364048, "rhcm_p90_m_s": 0.20574439734742442, "rham_p90_deg_s": null, "wms_p90": 0.9570865764220556, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "c521989bed067eb7f007eca7c67bac27f9e266b6baac75455d84c73afe0f30fb", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/7a92636dd7fa26f9c9cacade/reference.npz", "window_s": [ 0.0, 13.7 ], "timestamps_source": "source frame timeline", "frame_count": 412, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "356e51d450d870e0f211cc73c58c5f98aa1fb5498c23677186fcb1e8273a6691", "mano_sha256": "c8737e32967e2cccd79e47dffa90f4f8df2b2f4a8ed66b13544aee62543083ab", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "815ee42daf83ff85e272200f", "sample_id": "c9db05ce41b5523a45819435", "file": "20_seconds/H2O/c9db05ce41b5523a45819435_MANO_GT_20.00s.mp4", "recording_id": "subject3/h1/7", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "815ee42daf83ff85e272200f", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/c9db05ce41b5523a45819435_MANO_GT_20.00s.mp4", "bytes": 6647354, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.017033473872527728, "p90": 0.0327714541744986, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 1.7635416242707223, "p90": 3.835757392401489, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.020776567769380107, "p90": 0.10967530246443638, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01120740670552325, "p90": 0.02580730246749796, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.02469448388509543, "p90": 0.10641899783313918, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.024689381626022258, "p90": 0.10641139734632286, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.7585119234395468, "p90": 0.9381527427358244, "valid_pairs": 499, "valid_duration_s": 16.63333333333334, "coverage": 0.8330550918196999 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.25744122031702776, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.03909536314491047, "p90": 0.14895287924097844, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.011850261168698183, "p90": 0.027131631424548375, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.038837757048277056, "p90": 0.1439050166009796, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03882147228377772, "p90": 0.14391210713221012, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.783991700243778, "p90": 0.9417439331987307, "valid_pairs": 562, "valid_duration_s": 18.733333333333327, "coverage": 0.9382303839732886 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.32891469059357564, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0327714541744986, "camera_angular_speed_p90_deg_s": 3.835757392401489, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.14895287924097844, "hand_camera_speed_p90_m_s": 0.14391210713221012, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.027131631424548375, "rhcm_p90_m_s": 0.14391210713221012, "rham_p90_deg_s": null, "wms_p90": 0.9417439331987307, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "682a073361ea567dcb74521593b1aba0234c9419171f78cbbc614a1e40b2190f", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/815ee42daf83ff85e272200f/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7ace418842ac639254b3de24f14f6ea37c5d3f604d36ddfe5d3e0e56926539e5", "mano_sha256": "44f32a7399b4c658be6c5f2612b6c1a589fafe583c0bb30a7ed82a07072b78c3", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "87e0d7fecf4b103c2976427c", "sample_id": "4e13db653721021bdfb5d1a8", "file": "20_seconds/H2O/4e13db653721021bdfb5d1a8_MANO_GT_20.00s.mp4", "recording_id": "subject1/o2/6", "source_first_frame": "25", "source_last_frame": "624", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "87e0d7fecf4b103c2976427c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/4e13db653721021bdfb5d1a8_MANO_GT_20.00s.mp4", "bytes": 6220006, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02783185336875138, "p90": 0.06008941057691962, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.7480659600413846, "p90": 5.760671279864871, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03189098983173837, "p90": 0.11485225526283697, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01881493454471205, "p90": 0.04475882863919861, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.040917936834301505, "p90": 0.11682659640527886, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.04093715845490199, "p90": 0.11682488843121627, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.7707285488118806, "p90": 0.9392879387338643, "valid_pairs": 569, "valid_duration_s": 18.96666666666667, "coverage": 0.9499165275459098 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.10586613140071802, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07434087813393005, "p90": 0.2731537970817761, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0197416603597082, "p90": 0.04745079356105755, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.07360776690273224, "p90": 0.27590298669534136, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.07360886287464867, "p90": 0.2758980565830777, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8279705962934649, "p90": 0.9714467398621933, "valid_pairs": 595, "valid_duration_s": 19.833333333333332, "coverage": 0.9933222036727878 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.03492537730107592, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06008941057691962, "camera_angular_speed_p90_deg_s": 5.760671279864871, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2731537970817761, "hand_camera_speed_p90_m_s": 0.2758980565830777, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04745079356105755, "rhcm_p90_m_s": 0.2758980565830777, "rham_p90_deg_s": null, "wms_p90": 0.9714467398621933, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "8655ecfba3fbbbd9bc856256e5c6e4f4464306ce38132d7582652d4ac73b5294", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/87e0d7fecf4b103c2976427c/reference.npz", "window_s": [ 0.8333333333333334, 20.8 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "04ab7b7efd6d98e73a1833a3291237c2856e4eba2966517669a2f2a8d5c8a314", "mano_sha256": "a50ada877c809354c7266c10936de9dc3912caa4be8966f9844d98ece0e99675", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "91d6a5270adfca0ba8b42b2a", "sample_id": "16d82058091b572233e52752", "file": "20_seconds/H2O/16d82058091b572233e52752_MANO_GT_20.00s.mp4", "recording_id": "subject2/k1/1", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "91d6a5270adfca0ba8b42b2a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/16d82058091b572233e52752_MANO_GT_20.00s.mp4", "bytes": 7055288, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02496159245395201, "p90": 0.0628692180294198, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 2.764125807809229, "p90": 8.084809694476633, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07794140907611563, "p90": 0.3612299906558313, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01954142942751365, "p90": 0.05785172565388558, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08006858569137862, "p90": 0.3601359267346574, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.08005683827036257, "p90": 0.36017264413307587, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.8207062068296765, "p90": 0.96298253680789, "valid_pairs": 591, "valid_duration_s": 19.7, "coverage": 0.986644407345576 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.32052238680967715, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.044089891511009394, "p90": 0.18316145950209586, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.02014279443103926, "p90": 0.06967698397991909, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.04255981692454332, "p90": 0.15219480738154575, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.042560866940416796, "p90": 0.15225863478698684, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.6830006128920143, "p90": 0.9048846616943016, "valid_pairs": 555, "valid_duration_s": 18.499999999999993, "coverage": 0.9265442404006675 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.576358200121719, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0628692180294198, "camera_angular_speed_p90_deg_s": 8.084809694476633, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3612299906558313, "hand_camera_speed_p90_m_s": 0.36017264413307587, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.06967698397991909, "rhcm_p90_m_s": 0.36017264413307587, "rham_p90_deg_s": null, "wms_p90": 0.96298253680789, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "8fd676d9a66fb8d0b70ecfb7b00c8a0f4a70c722277533a7c5cb107f5fe6a0aa", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/91d6a5270adfca0ba8b42b2a/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8b43f117e99856e13d4ea957bfbcb6a97e2a44a60bdae0186a093dead7181318", "mano_sha256": "8c3f7c1054012a580c9392762129287cdc38d4a5d5bcdfb388017d56b62e242e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "9311161e42461a2750697b12", "sample_id": "8814eee966c01ea8c48cda7b", "file": "20_seconds/H2O/8814eee966c01ea8c48cda7b_MANO_GT_20.00s.mp4", "recording_id": "subject3/k1/2", "source_first_frame": "161", "source_last_frame": "760", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "9311161e42461a2750697b12", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/8814eee966c01ea8c48cda7b_MANO_GT_20.00s.mp4", "bytes": 8565302, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02325512345537522, "p90": 0.039730691998415404, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 2.695671757860705, "p90": 5.707289936273164, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.04915158573981798, "p90": 0.1328121886074137, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.01925003541503169, "p90": 0.04056762049432493, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.05441496907710299, "p90": 0.13383503675485917, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.05440151632062894, "p90": 0.1338351474463998, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.79666363425949, "p90": 0.9566235449668506, "valid_pairs": 594, "valid_duration_s": 19.800000000000004, "coverage": 0.9916527545909852 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.009388110619147853, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.05823842886520336, "p90": 0.16830214843244123, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.02015534731967089, "p90": 0.039578372322569264, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.06367531774894418, "p90": 0.17329929026766397, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.06368541072954864, "p90": 0.17329227085780963, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.8032580976814814, "p90": 0.9567915520096302, "valid_pairs": 596, "valid_duration_s": 19.866666666666667, "coverage": 0.9949916527545911 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.04597394766081707, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.039730691998415404, "camera_angular_speed_p90_deg_s": 5.707289936273164, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.16830214843244123, "hand_camera_speed_p90_m_s": 0.17329227085780963, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04056762049432493, "rhcm_p90_m_s": 0.17329227085780963, "rham_p90_deg_s": null, "wms_p90": 0.9567915520096302, "gt_coverage": 1.0000000000000002, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "b07c34bffe8cbde2b285f9c7d429d413750300a282052f93d05b324334c4f3d3", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/9311161e42461a2750697b12/reference.npz", "window_s": [ 5.366666666666666, 25.333333333333332 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b0c64c711131f87769c63aa9c5377e0d70dc06d6dafd3434f847009ce3168204", "mano_sha256": "5217c2758ca5c7e68d2a078ceda78dcff97fa74ec98fab8c795237817e8d3c94", "duration_s": 14.634, "frames": 439, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "9d353aae7feb1d0912e12b0f", "sample_id": "c1286fd19d0f2ae657ad6494", "file": "short_originals_under_20s/H2O/c1286fd19d0f2ae657ad6494_MANO_GT_14.63s.mp4", "recording_id": "subject3/h1/4", "source_first_frame": "0", "source_last_frame": "438", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "9d353aae7feb1d0912e12b0f", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/c1286fd19d0f2ae657ad6494_MANO_GT_14.63s.mp4", "bytes": 5363288, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.025310611039725307, "p90": 0.047464836001460405, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 1.0 }, "camera_rotation": { "median": 2.9650096577552407, "p90": 6.401131449874974, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.030161215878822267, "p90": 0.13906306508941085, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01870308036469593, "p90": 0.039401098216985946, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03895622330825872, "p90": 0.13164477901985602, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03893230501559296, "p90": 0.13164158376822635, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 1.0 }, "wms": { "median": 0.7523820512121703, "p90": 0.9326658834561147, "valid_pairs": 414, "valid_duration_s": 13.8, "coverage": 0.9452054794520549 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.19025524510303407, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07819805911185751, "p90": 0.29886765455341735, "valid_pairs": 391, "valid_duration_s": 13.033333333333333, "coverage": 0.8926940639269406 }, "cmh_m_s": { "median": 0.01893511705129056, "p90": 0.035522979491480165, "valid_pairs": 391, "valid_duration_s": 13.033333333333333, "coverage": 0.8926940639269406 }, "rhcm_continuous_estimate_m_s": { "median": 0.075314872553874, "p90": 0.30629826210484146, "valid_pairs": 391, "valid_duration_s": 13.033333333333333, "coverage": 0.8926940639269406 }, "rhcm_discrete_m_s": { "median": 0.07534909123962498, "p90": 0.3062988066975501, "valid_pairs": 391, "valid_duration_s": 13.033333333333333, "coverage": 0.8926940639269406 }, "wms": { "median": 0.8404744002470139, "p90": 0.9652100301850133, "valid_pairs": 382, "valid_duration_s": 12.733333333333334, "coverage": 0.8721461187214613 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.2472093854264114, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.047464836001460405, "camera_angular_speed_p90_deg_s": 6.401131449874974, "camera_valid_fraction": 1.0, "camera_pairs": 438, "hand_world_speed_p90_m_s": 0.29886765455341735, "hand_camera_speed_p90_m_s": 0.3062988066975501, "hand_world_valid_fraction": 0.9476082004555809, "hand_camera_valid_fraction": 0.9476082004555809, "cmh_p90_m_s": 0.039401098216985946, "rhcm_p90_m_s": 0.3062988066975501, "rham_p90_deg_s": null, "wms_p90": 0.9652100301850133, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "3bde2b3eac09ef69500d37de48325fbb50a32fc13e12632a83617e42674b0bc8", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/9d353aae7feb1d0912e12b0f/reference.npz", "window_s": [ 0.0, 14.6 ], "timestamps_source": "source frame timeline", "frame_count": 439, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3569462ec676454a943799482934f680cf5243a388a917aabc536d77d0207b73", "mano_sha256": "1c19efc508b706ba2401a852160d51623342252a1f4139720efa51235a18a7b1", "duration_s": 15.9, "frames": 477, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "9d47c7b397974a4dddd690d2", "sample_id": "0f84b5933cf28f0cced5af90", "file": "short_originals_under_20s/H2O/0f84b5933cf28f0cced5af90_MANO_GT_15.90s.mp4", "recording_id": "subject3/h2/4", "source_first_frame": "0", "source_last_frame": "476", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "9d47c7b397974a4dddd690d2", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/0f84b5933cf28f0cced5af90_MANO_GT_15.90s.mp4", "bytes": 6537139, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.032920037453193926, "p90": 0.07298156805578498, "valid_pairs": 476, "valid_duration_s": 15.866666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.296022466767136, "p90": 7.968378610692643, "valid_pairs": 476, "valid_duration_s": 15.866666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.05025477852832848, "p90": 0.21448536071329904, "valid_pairs": 476, "valid_duration_s": 15.866666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.023557720178947243, "p90": 0.05921337003718148, "valid_pairs": 476, "valid_duration_s": 15.866666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.055930275805218505, "p90": 0.19945939622081216, "valid_pairs": 476, "valid_duration_s": 15.866666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.05584711131381641, "p90": 0.19946813618855275, "valid_pairs": 476, "valid_duration_s": 15.866666666666667, "coverage": 1.0 }, "wms": { "median": 0.7732146391654374, "p90": 0.9374033628385984, "valid_pairs": 464, "valid_duration_s": 15.466666666666665, "coverage": 0.9747899159663864 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4174032194804436, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08704789860310937, "p90": 0.30065571728142204, "valid_pairs": 441, "valid_duration_s": 14.7, "coverage": 0.926470588235294 }, "cmh_m_s": { "median": 0.022653990115207665, "p90": 0.05607384054714259, "valid_pairs": 441, "valid_duration_s": 14.7, "coverage": 0.926470588235294 }, "rhcm_continuous_estimate_m_s": { "median": 0.0880770268574811, "p90": 0.30503454568433686, "valid_pairs": 441, "valid_duration_s": 14.7, "coverage": 0.926470588235294 }, "rhcm_discrete_m_s": { "median": 0.08804561256134559, "p90": 0.3050820006731197, "valid_pairs": 441, "valid_duration_s": 14.7, "coverage": 0.926470588235294 }, "wms": { "median": 0.8009172776015261, "p90": 0.9412640395852037, "valid_pairs": 435, "valid_duration_s": 14.5, "coverage": 0.9138655462184874 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4882571210302303, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.07298156805578498, "camera_angular_speed_p90_deg_s": 7.968378610692643, "camera_valid_fraction": 1.0, "camera_pairs": 476, "hand_world_speed_p90_m_s": 0.30065571728142204, "hand_camera_speed_p90_m_s": 0.3050820006731197, "hand_world_valid_fraction": 0.9643605870020965, "hand_camera_valid_fraction": 0.9643605870020965, "cmh_p90_m_s": 0.05921337003718148, "rhcm_p90_m_s": 0.3050820006731197, "rham_p90_deg_s": null, "wms_p90": 0.9412640395852037, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "c4640e081b3eb84df175145a26ce0c674434e73e4177c53e597ae6b157d20951", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/9d47c7b397974a4dddd690d2/reference.npz", "window_s": [ 0.0, 15.866666666666667 ], "timestamps_source": "source frame timeline", "frame_count": 477, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b1975d4889b59c4bc7fd44ba7fd7ecda8d320bcf024348f493debf534ce0fb1f", "mano_sha256": "7b38f2be522cac945e23fec909c422ba2e421fc4951eaf008d7bda3a59008b5a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a0cfda25aa636427f1483d8b", "sample_id": "7a01e37cfc9e8ee671cebcad", "file": "20_seconds/H2O/7a01e37cfc9e8ee671cebcad_MANO_GT_20.00s.mp4", "recording_id": "subject2/h2/3", "source_first_frame": "211", "source_last_frame": "810", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "a0cfda25aa636427f1483d8b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/7a01e37cfc9e8ee671cebcad_MANO_GT_20.00s.mp4", "bytes": 7250584, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.019610287062431186, "p90": 0.040265293385349774, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 1.8871872160700678, "p90": 4.595971749862595, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08308515480575515, "p90": 0.18791593333864154, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.013484670917078317, "p90": 0.031169320879749964, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08054025175973072, "p90": 0.1863091427518425, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.08055518361020325, "p90": 0.18630902907728905, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8594419481090458, "p90": 0.9645308950489921, "valid_pairs": 590, "valid_duration_s": 19.666666666666664, "coverage": 0.9849749582637728 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.05517439880291008, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.038693104051008155, "p90": 0.12420695525118729, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.013723092676649193, "p90": 0.03303567790364062, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.0441040509151161, "p90": 0.12892287202299307, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.04409661816996161, "p90": 0.1289214728610633, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.8048260529561572, "p90": 0.9505126168616158, "valid_pairs": 554, "valid_duration_s": 18.466666666666672, "coverage": 0.924874791318865 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.33480790168675767, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.040265293385349774, "camera_angular_speed_p90_deg_s": 4.595971749862595, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.18791593333864154, "hand_camera_speed_p90_m_s": 0.18630902907728905, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.03303567790364062, "rhcm_p90_m_s": 0.18630902907728905, "rham_p90_deg_s": null, "wms_p90": 0.9645308950489921, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "e849daee12e946c94caa8d1686e66641008640aa5641f357d27f35fcb6cb0a92", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/a0cfda25aa636427f1483d8b/reference.npz", "window_s": [ 7.033333333333333, 27.0 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b5740f84f0a1cb8c4b1c58e177beadea48cf49396dfcebd51bfd35281ed18b5e", "mano_sha256": "a2649c1a724bcc796a4e92c97339f25a415a7e063b054c84849020d6ccb3705f", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a64a57f08b734b430fdb41a0", "sample_id": "ed40cbf7ed5261fa43ade76f", "file": "20_seconds/H2O/ed40cbf7ed5261fa43ade76f_MANO_GT_20.00s.mp4", "recording_id": "subject1/k1/7", "source_first_frame": "2", "source_last_frame": "601", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "a64a57f08b734b430fdb41a0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/ed40cbf7ed5261fa43ade76f_MANO_GT_20.00s.mp4", "bytes": 6367060, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02735888164512633, "p90": 0.08077457721465849, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.943425149216821, "p90": 10.252637361300529, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.028954554339869463, "p90": 0.10485389647486967, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01861595895095148, "p90": 0.06445915388783532, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03826505693121607, "p90": 0.10971543247888974, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03827592589741869, "p90": 0.10975687598264078, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.7582210111460083, "p90": 0.9371629258609311, "valid_pairs": 576, "valid_duration_s": 19.200000000000003, "coverage": 0.9616026711185309 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.19568344118681635, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.05335435432336951, "p90": 0.24353819788109016, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.02125328937778275, "p90": 0.06305025068577523, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06431131099061396, "p90": 0.21602652518135312, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0643146495471329, "p90": 0.21620437320314836, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.7923505919062983, "p90": 0.9477566334386175, "valid_pairs": 594, "valid_duration_s": 19.8, "coverage": 0.9916527545909849 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.21586585222863075, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.08077457721465849, "camera_angular_speed_p90_deg_s": 10.252637361300529, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.24353819788109016, "hand_camera_speed_p90_m_s": 0.21620437320314836, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.06445915388783532, "rhcm_p90_m_s": 0.21620437320314836, "rham_p90_deg_s": null, "wms_p90": 0.9477566334386175, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "b9d40a8f1074c6d934b828243f65a4728588aee2afe0b22e3b6f2f50ecec8ff4", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/a64a57f08b734b430fdb41a0/reference.npz", "window_s": [ 0.06666666666666667, 20.033333333333335 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6f9738c2f901a034c788a0bacc1f0f36d372f76b5c159367e6cbf55ce3d2a4b1", "mano_sha256": "f8126fbd2642a8696820c97f29b7aecada7a9cc479b30b82160c848603285003", "duration_s": 19.8, "frames": 594, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a66fd4d66c45a8faa8164858", "sample_id": "0da2afd5fdd0f6c2a2897c24", "file": "short_originals_under_20s/H2O/0da2afd5fdd0f6c2a2897c24_MANO_GT_19.80s.mp4", "recording_id": "subject1/k2/5", "source_first_frame": "0", "source_last_frame": "593", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "a66fd4d66c45a8faa8164858", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/0da2afd5fdd0f6c2a2897c24_MANO_GT_19.80s.mp4", "bytes": 7175113, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.025241290522854197, "p90": 0.05934441541308137, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "camera_rotation": { "median": 2.6118290365154038, "p90": 5.785371762235815, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03232006490403659, "p90": 0.11623520472112019, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01781346198537697, "p90": 0.03793668519651223, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03932498152683895, "p90": 0.11023812527746851, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03929752038141303, "p90": 0.11015535264806287, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "wms": { "median": 0.7690805884219891, "p90": 0.9464139056375807, "valid_pairs": 550, "valid_duration_s": 18.333333333333325, "coverage": 0.9274873524451935 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.14655171421613244, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07798347604641968, "p90": 0.2468130748382987, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.017811640481309687, "p90": 0.04089765063415151, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.07717850887818824, "p90": 0.23757325376129618, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.07723284265686955, "p90": 0.23758645316362992, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 1.0 }, "wms": { "median": 0.8215362006574266, "p90": 0.955999428577733, "valid_pairs": 588, "valid_duration_s": 19.599999999999998, "coverage": 0.9915682967959527 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.295265390332844, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05934441541308137, "camera_angular_speed_p90_deg_s": 5.785371762235815, "camera_valid_fraction": 1.0, "camera_pairs": 593, "hand_world_speed_p90_m_s": 0.2468130748382987, "hand_camera_speed_p90_m_s": 0.23758645316362992, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04089765063415151, "rhcm_p90_m_s": 0.23758645316362992, "rham_p90_deg_s": null, "wms_p90": 0.955999428577733, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "931069d626476f8425df8a5bb27a18e8bb02607ae1009c19b508878ea3db2b5c", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/a66fd4d66c45a8faa8164858/reference.npz", "window_s": [ 0.0, 19.766666666666666 ], "timestamps_source": "source frame timeline", "frame_count": 594, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ad668f2614e9723926c6a6890aba39b687bedc049837c6f777b01584f9bf0f7f", "mano_sha256": "8b389956fc8e7a2bcbc6e22606446737a2608ca8d6d37d9ae5226a310c2fb30a", "duration_s": 13.334, "frames": 400, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a8f0bbe0ca61fb97582e4470", "sample_id": "5ccb31a7bcd7f99c81c3400c", "file": "short_originals_under_20s/H2O/5ccb31a7bcd7f99c81c3400c_MANO_GT_13.33s.mp4", "recording_id": "subject1/h2/0", "source_first_frame": "0", "source_last_frame": "399", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "a8f0bbe0ca61fb97582e4470", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/5ccb31a7bcd7f99c81c3400c_MANO_GT_13.33s.mp4", "bytes": 5404310, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03706158993401758, "p90": 0.11692550661915165, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "camera_rotation": { "median": 4.6005926087408575, "p90": 13.106125290083833, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06193766441391466, "p90": 0.33428187527629566, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "cmh_m_s": { "median": 0.029432067403871456, "p90": 0.06729454745423454, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06827735454721788, "p90": 0.32321621422426144, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0682775517027121, "p90": 0.3232054860145351, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "wms": { "median": 0.805591229997887, "p90": 0.963652275613266, "valid_pairs": 395, "valid_duration_s": 13.166666666666668, "coverage": 0.9899749373433584 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": -0.01599652397324971, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.027725250371843525, "p90": 0.16447378621465006, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "cmh_m_s": { "median": 0.03062374853156887, "p90": 0.09033320596485489, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.04315023845750439, "p90": 0.15312889844810337, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.043148158792169725, "p90": 0.15319347916255135, "valid_pairs": 399, "valid_duration_s": 13.3, "coverage": 1.0 }, "wms": { "median": 0.7247333134059725, "p90": 0.9305264016021908, "valid_pairs": 380, "valid_duration_s": 12.666666666666668, "coverage": 0.9523809523809524 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4723196811123285, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.11692550661915165, "camera_angular_speed_p90_deg_s": 13.106125290083833, "camera_valid_fraction": 1.0, "camera_pairs": 399, "hand_world_speed_p90_m_s": 0.33428187527629566, "hand_camera_speed_p90_m_s": 0.3232054860145351, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.09033320596485489, "rhcm_p90_m_s": 0.3232054860145351, "rham_p90_deg_s": null, "wms_p90": 0.963652275613266, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "0cba977d0ff1d3dc0d301bc50b8f8b266947a67b5244ffa6c7eadaff2f261819", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/a8f0bbe0ca61fb97582e4470/reference.npz", "window_s": [ 0.0, 13.3 ], "timestamps_source": "source frame timeline", "frame_count": 400, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3284086f2c113961739d493084d1416ff2a8bbb8f5b1095d18d33c78777d2b98", "mano_sha256": "6de1897734d9b26d7aa6957a5c145686e2d8e1bbd32839d571c7a4f409676ae4", "duration_s": 17.567, "frames": 527, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "aae395e60e522fd0f7a9848e", "sample_id": "88e9073eedeaf2040a6d4257", "file": "short_originals_under_20s/H2O/88e9073eedeaf2040a6d4257_MANO_GT_17.57s.mp4", "recording_id": "subject2/o1/0", "source_first_frame": "0", "source_last_frame": "526", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "aae395e60e522fd0f7a9848e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/88e9073eedeaf2040a6d4257_MANO_GT_17.57s.mp4", "bytes": 5277873, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.023129910849795578, "p90": 0.05103912410368783, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "camera_rotation": { "median": 2.4983745930194923, "p90": 5.1146949458594015, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.057442802231744, "p90": 0.26863053543320903, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01934854395188225, "p90": 0.04080945012054685, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05923873297107614, "p90": 0.26464629401870576, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.05925310632671789, "p90": 0.26464521054727963, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "wms": { "median": 0.8147315941998432, "p90": 0.9592772360522254, "valid_pairs": 509, "valid_duration_s": 16.96666666666667, "coverage": 0.967680608365019 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.30487683556189554, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07955041683191424, "p90": 0.2415230092897252, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "cmh_m_s": { "median": 0.019772874821065697, "p90": 0.04218707722531309, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.079261951053301, "p90": 0.24116586924209493, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.07926280322545269, "p90": 0.24117606641640232, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 1.0 }, "wms": { "median": 0.8183265099024959, "p90": 0.9550006224952884, "valid_pairs": 525, "valid_duration_s": 17.5, "coverage": 0.9980988593155893 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.2272186038649484, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05103912410368783, "camera_angular_speed_p90_deg_s": 5.1146949458594015, "camera_valid_fraction": 1.0, "camera_pairs": 526, "hand_world_speed_p90_m_s": 0.26863053543320903, "hand_camera_speed_p90_m_s": 0.26464521054727963, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04218707722531309, "rhcm_p90_m_s": 0.26464521054727963, "rham_p90_deg_s": null, "wms_p90": 0.9592772360522254, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "2a1aaff5d89bd25353a1eaacae43cbd14bf82683c55278cb6998a986e16395a1", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/aae395e60e522fd0f7a9848e/reference.npz", "window_s": [ 0.0, 17.533333333333335 ], "timestamps_source": "source frame timeline", "frame_count": 527, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "390e26455f94e4ceff748992d41a6fe39808624218931d246033e1388e7c227f", "mano_sha256": "2880e3d2a7674f0166a0a1f8fc891fc7185d2f146025d9a53dc3300d70651573", "duration_s": 15.734, "frames": 472, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "官方 MANO 覆盖双手，但指部与物体接触仍有近似偏差；不能自动认为标签准确。", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "b50c8a8fd4fab0d471c666ea", "sample_id": "17d87645af1ca06eb502c3c3", "file": "short_originals_under_20s/H2O/17d87645af1ca06eb502c3c3_MANO_GT_15.73s.mp4", "recording_id": "subject1/k1/6", "source_first_frame": "0", "source_last_frame": "471", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "b50c8a8fd4fab0d471c666ea", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/17d87645af1ca06eb502c3c3_MANO_GT_15.73s.mp4", "bytes": 5007773, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02467527665745763, "p90": 0.050513421225935065, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "camera_rotation": { "median": 2.814226915801392, "p90": 5.789756740504649, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0326776353930773, "p90": 0.1295042912141306, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01924601986195486, "p90": 0.03751361864748967, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03999467533027402, "p90": 0.127009359286037, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0399367640123602, "p90": 0.12701499548866693, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "wms": { "median": 0.757125434785332, "p90": 0.9380197040549509, "valid_pairs": 438, "valid_duration_s": 14.599999999999998, "coverage": 0.929936305732484 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.27949850813462856, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07170118128332259, "p90": 0.26827192879939693, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01948009286840676, "p90": 0.040154641982945785, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.07181765179638193, "p90": 0.26140327170196914, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.07182123820104536, "p90": 0.2614308938584817, "valid_pairs": 471, "valid_duration_s": 15.7, "coverage": 1.0 }, "wms": { "median": 0.8150596674112174, "p90": 0.9588575671505845, "valid_pairs": 466, "valid_duration_s": 15.533333333333331, "coverage": 0.989384288747346 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.19933836517281064, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.050513421225935065, "camera_angular_speed_p90_deg_s": 5.789756740504649, "camera_valid_fraction": 1.0, "camera_pairs": 471, "hand_world_speed_p90_m_s": 0.26827192879939693, "hand_camera_speed_p90_m_s": 0.2614308938584817, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.040154641982945785, "rhcm_p90_m_s": 0.2614308938584817, "rham_p90_deg_s": null, "wms_p90": 0.9588575671505845, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "9adb60b3a0f7b138674a862035e4fe8b58423015849378cecebd5befd8660cce", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/b50c8a8fd4fab0d471c666ea/reference.npz", "window_s": [ 0.0, 15.7 ], "timestamps_source": "source frame timeline", "frame_count": 472, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "18e19bad1e526f600f3dd65a2fb44c1f90ffb66f3f6fbb01018868a177a0ffad", "mano_sha256": "cf4cf979e9c6296df2f744eedc7107bcf62d0071ac49dbba08e30f547f8bae6f", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "b9eef1e8d652a63ad26f2f95", "sample_id": "f76d18e3b8bfbce55febd92f", "file": "20_seconds/H2O/f76d18e3b8bfbce55febd92f_MANO_GT_20.00s.mp4", "recording_id": "subject1/o1/3", "source_first_frame": "64", "source_last_frame": "663", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "b9eef1e8d652a63ad26f2f95", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/f76d18e3b8bfbce55febd92f_MANO_GT_20.00s.mp4", "bytes": 7093107, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03679488873746979, "p90": 0.07197731941960624, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.9877879589695926, "p90": 7.658504949881369, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.04196708836322359, "p90": 0.12486237676272904, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.02968747445983746, "p90": 0.05839962039749274, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.04731568798285293, "p90": 0.11668596975795223, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.047313564864521276, "p90": 0.11673985691743148, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.6911826784580037, "p90": 0.9262381631906107, "valid_pairs": 590, "valid_duration_s": 19.666666666666668, "coverage": 0.9849749582637729 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3514919514687243, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09926438585280702, "p90": 0.37287996349574687, "valid_pairs": 587, "valid_duration_s": 19.566666666666666, "coverage": 0.9799666110183638 }, "cmh_m_s": { "median": 0.027331575676971837, "p90": 0.05422487327478131, "valid_pairs": 587, "valid_duration_s": 19.566666666666666, "coverage": 0.9799666110183638 }, "rhcm_continuous_estimate_m_s": { "median": 0.10525834181848824, "p90": 0.37819510972008724, "valid_pairs": 587, "valid_duration_s": 19.566666666666666, "coverage": 0.9799666110183638 }, "rhcm_discrete_m_s": { "median": 0.1052981093346345, "p90": 0.3781957585617916, "valid_pairs": 587, "valid_duration_s": 19.566666666666666, "coverage": 0.9799666110183638 }, "wms": { "median": 0.8622307074238107, "p90": 0.969870328729713, "valid_pairs": 586, "valid_duration_s": 19.53333333333333, "coverage": 0.9782971619365608 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.02173239012625127, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.07197731941960624, "camera_angular_speed_p90_deg_s": 7.658504949881369, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.37287996349574687, "hand_camera_speed_p90_m_s": 0.3781957585617916, "hand_world_valid_fraction": 0.9916666666666667, "hand_camera_valid_fraction": 0.9916666666666667, "cmh_p90_m_s": 0.05839962039749274, "rhcm_p90_m_s": 0.3781957585617916, "rham_p90_deg_s": null, "wms_p90": 0.969870328729713, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "bc88b50ef52d34a565f5cea6086c08e3022d667cd69d1fe8be7afbb11349c1ad", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/b9eef1e8d652a63ad26f2f95/reference.npz", "window_s": [ 2.1333333333333333, 22.1 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a4a21bf0f55f20aeaab14f0b59a131f9094c06aa50806b37823e561449401a6b", "mano_sha256": "82fa356f6d964f80519ccc82556d8dbb012f28c1ae52ad629c34811cb04f1fe5", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "bc084e01837a0904614953d1", "sample_id": "e7309f21f47466c1460d4fd3", "file": "20_seconds/H2O/e7309f21f47466c1460d4fd3_MANO_GT_20.00s.mp4", "recording_id": "subject1/o1/7", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "bc084e01837a0904614953d1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/e7309f21f47466c1460d4fd3_MANO_GT_20.00s.mp4", "bytes": 6884397, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03592329723290309, "p90": 0.09559038373806802, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 3.082846054752035, "p90": 9.855014072909412, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.019855720783063812, "p90": 0.12140585479581757, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.02408316533522699, "p90": 0.08541881525639614, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.03481080788183033, "p90": 0.11411809495654206, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.034774258556626775, "p90": 0.11410170527571482, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.7580393045301218, "p90": 0.952685479536547, "valid_pairs": 496, "valid_duration_s": 16.533333333333335, "coverage": 0.8280467445742906 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.38441253817678284, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.049738919934176236, "p90": 0.22524339787180503, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.024020633221710417, "p90": 0.0739308322729781, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.05701494533127303, "p90": 0.2106745711501738, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.05699851521882306, "p90": 0.21061101328029772, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.7738904736967851, "p90": 0.9559969124126074, "valid_pairs": 593, "valid_duration_s": 19.766666666666666, "coverage": 0.989983305509182 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.10415530901558338, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.09559038373806802, "camera_angular_speed_p90_deg_s": 9.855014072909412, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.22524339787180503, "hand_camera_speed_p90_m_s": 0.21061101328029772, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.08541881525639614, "rhcm_p90_m_s": 0.21061101328029772, "rham_p90_deg_s": null, "wms_p90": 0.9559969124126074, "gt_coverage": 1.0000000000000002, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "7db77e121a18c6ee55bd9906b8fd4eee298ab14db78a2ca535593df94d03f8e1", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/bc084e01837a0904614953d1/reference.npz", "window_s": [ 0.03333333333333333, 20.0 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7a40bc5376b98f8e2cf8f5613258b3eeddb57e3c4198e50fc73750a661dc9b39", "mano_sha256": "fef13c4c9e44354b1dea7d3700ddd618270be416f669acb66d4a8d210351190a", "duration_s": 17.7, "frames": 531, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "bff0ea8811f24092c170f377", "sample_id": "cc124099ef04e0e4575d8e71", "file": "short_originals_under_20s/H2O/cc124099ef04e0e4575d8e71_MANO_GT_17.70s.mp4", "recording_id": "subject2/o2/4", "source_first_frame": "0", "source_last_frame": "530", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "bff0ea8811f24092c170f377", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/cc124099ef04e0e4575d8e71_MANO_GT_17.70s.mp4", "bytes": 5754186, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.026763191754589227, "p90": 0.05688291294512987, "valid_pairs": 530, "valid_duration_s": 17.666666666666668, "coverage": 1.0 }, "camera_rotation": { "median": 3.5849617165936616, "p90": 6.931098229721677, "valid_pairs": 530, "valid_duration_s": 17.666666666666668, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08951075039659266, "p90": 0.24083115553650591, "valid_pairs": 530, "valid_duration_s": 17.666666666666668, "coverage": 1.0 }, "cmh_m_s": { "median": 0.029507534145786517, "p90": 0.06707044559533018, "valid_pairs": 530, "valid_duration_s": 17.666666666666668, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.09267316881324575, "p90": 0.25335657363222636, "valid_pairs": 530, "valid_duration_s": 17.666666666666668, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.09267814444803232, "p90": 0.2533835648083683, "valid_pairs": 530, "valid_duration_s": 17.666666666666668, "coverage": 1.0 }, "wms": { "median": 0.7948816550336832, "p90": 0.9635304800422404, "valid_pairs": 526, "valid_duration_s": 17.53333333333333, "coverage": 0.992452830188679 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3844136148580095, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.14797880458454693, "p90": 0.29394074259162095, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.9849056603773585 }, "cmh_m_s": { "median": 0.02769987558096449, "p90": 0.058504666826647504, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.9849056603773585 }, "rhcm_continuous_estimate_m_s": { "median": 0.13873896525399593, "p90": 0.307583405255624, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.9849056603773585 }, "rhcm_discrete_m_s": { "median": 0.13873998043493918, "p90": 0.30757415262834636, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.9849056603773585 }, "wms": { "median": 0.8353839287505496, "p90": 0.9523135454388071, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.9830188679245283 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.45511733074057437, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05688291294512987, "camera_angular_speed_p90_deg_s": 6.931098229721677, "camera_valid_fraction": 1.0, "camera_pairs": 530, "hand_world_speed_p90_m_s": 0.29394074259162095, "hand_camera_speed_p90_m_s": 0.30757415262834636, "hand_world_valid_fraction": 0.9943502824858758, "hand_camera_valid_fraction": 0.9943502824858758, "cmh_p90_m_s": 0.06707044559533018, "rhcm_p90_m_s": 0.30757415262834636, "rham_p90_deg_s": null, "wms_p90": 0.9635304800422404, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "b54eb15e7fc86bb1770a4cc6597219f979442c0a764967f1f7c7727cbd7c1ca5", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/bff0ea8811f24092c170f377/reference.npz", "window_s": [ 0.0, 17.666666666666668 ], "timestamps_source": "source frame timeline", "frame_count": 531, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "aaf4f928092933d4266c16c8a35ca63597ead8e868ad9a29b90d3aaa5d016c24", "mano_sha256": "b16c7ecf18dcf0213e62080c59c31d78d7da6a0c5cf8bfed8b042587d16c6021", "duration_s": 15.8, "frames": 474, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "ccc1917d48fc1143cfe9a9a0", "sample_id": "cb9ee1bedc15a2aea8316093", "file": "short_originals_under_20s/H2O/cb9ee1bedc15a2aea8316093_MANO_GT_15.80s.mp4", "recording_id": "subject2/k1/3", "source_first_frame": "0", "source_last_frame": "473", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "ccc1917d48fc1143cfe9a9a0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/cb9ee1bedc15a2aea8316093_MANO_GT_15.80s.mp4", "bytes": 5361614, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.022438415414091712, "p90": 0.04349464940510368, "valid_pairs": 473, "valid_duration_s": 15.76666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 2.8508526737635838, "p90": 5.3508800248834705, "valid_pairs": 473, "valid_duration_s": 15.76666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.043027480822515875, "p90": 0.2200728372293296, "valid_pairs": 473, "valid_duration_s": 15.76666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.023770536839267054, "p90": 0.048512477735957904, "valid_pairs": 473, "valid_duration_s": 15.76666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.05255760108323969, "p90": 0.21701247810546515, "valid_pairs": 473, "valid_duration_s": 15.76666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.052564386468661184, "p90": 0.21700770079667184, "valid_pairs": 473, "valid_duration_s": 15.76666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.7666614324225286, "p90": 0.9380659734712544, "valid_pairs": 467, "valid_duration_s": 15.566666666666665, "coverage": 0.9873150105708244 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.033569114958226995, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.11315419884611305, "p90": 0.3640891183015859, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.9365750528541226 }, "cmh_m_s": { "median": 0.021680429904508403, "p90": 0.044936982960018725, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.9365750528541226 }, "rhcm_continuous_estimate_m_s": { "median": 0.10952444644299665, "p90": 0.35727922725878, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.9365750528541226 }, "rhcm_discrete_m_s": { "median": 0.10950221958607692, "p90": 0.3572839373881447, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.9365750528541226 }, "wms": { "median": 0.8403704851369262, "p90": 0.9707716369540766, "valid_pairs": 442, "valid_duration_s": 14.733333333333334, "coverage": 0.9344608879492601 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.20885551087383517, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04349464940510368, "camera_angular_speed_p90_deg_s": 5.3508800248834705, "camera_valid_fraction": 1.0, "camera_pairs": 473, "hand_world_speed_p90_m_s": 0.3640891183015859, "hand_camera_speed_p90_m_s": 0.3572839373881447, "hand_world_valid_fraction": 0.9694092827004219, "hand_camera_valid_fraction": 0.9694092827004219, "cmh_p90_m_s": 0.048512477735957904, "rhcm_p90_m_s": 0.3572839373881447, "rham_p90_deg_s": null, "wms_p90": 0.9707716369540766, "gt_coverage": 1.0000000000000002, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "625ec9564a506557266caea3d0f921e4c7ce38558dc9854c1ba1c1b1224c89a9", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/ccc1917d48fc1143cfe9a9a0/reference.npz", "window_s": [ 0.0, 15.766666666666667 ], "timestamps_source": "source frame timeline", "frame_count": 474, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "138dbaf9f650e44e1b9c1857e1a2f26c459f3c11f8c5ee09af22fe71c7e972ca", "mano_sha256": "5a15df115ee7a44169e52239c96c8695896ee951141983a7c4b5a8e256f0cfaa", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "d056c73e829b8cb2331ac216", "sample_id": "194e6b92d8e2dfb0b2f6dc76", "file": "20_seconds/H2O/194e6b92d8e2dfb0b2f6dc76_MANO_GT_20.00s.mp4", "recording_id": "subject2/k1/6", "source_first_frame": "232", "source_last_frame": "831", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "d056c73e829b8cb2331ac216", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/194e6b92d8e2dfb0b2f6dc76_MANO_GT_20.00s.mp4", "bytes": 6481008, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.01777994863843234, "p90": 0.03226040826781424, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 1.770690260649402, "p90": 3.3474273795010743, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.043205691288001384, "p90": 0.1227190137324734, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01327841514213149, "p90": 0.026966449041646398, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.04398137784533169, "p90": 0.12368578623301996, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.043995476207659125, "p90": 0.12367906939226414, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.8230137844976798, "p90": 0.9517233615446655, "valid_pairs": 579, "valid_duration_s": 19.299999999999997, "coverage": 0.9666110183639398 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.08609901675590867, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.04653137475839726, "p90": 0.1326083185347763, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "cmh_m_s": { "median": 0.013406485584252368, "p90": 0.02822051397395765, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05358339496442031, "p90": 0.13080443604439057, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.053586184364049855, "p90": 0.13080437359130437, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "wms": { "median": 0.838290710854825, "p90": 0.9620700911608594, "valid_pairs": 578, "valid_duration_s": 19.26666666666667, "coverage": 0.9649415692821371 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.15889900112227176, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.03226040826781424, "camera_angular_speed_p90_deg_s": 3.3474273795010743, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.1326083185347763, "hand_camera_speed_p90_m_s": 0.13080437359130437, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.02822051397395765, "rhcm_p90_m_s": 0.13080437359130437, "rham_p90_deg_s": null, "wms_p90": 0.9620700911608594, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "17452ac4ec8b06d1ae306451ec1e9bfa13cf888185fbe24123b42345f2ce70ab", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/d056c73e829b8cb2331ac216/reference.npz", "window_s": [ 7.733333333333333, 27.7 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6ba1d95164fbb8d9ee618f18621be6396c6883e89de503b20940f9166efd3c57", "mano_sha256": "bebaa408b143e2c19da856e3ac3a0c9092d14cd6499bf001d5137c1e290ab576", "duration_s": 14.9, "frames": 447, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "d0fbe97e1d72758881a3e4f2", "sample_id": "b57bc8e5cacc8da53a29b024", "file": "short_originals_under_20s/H2O/b57bc8e5cacc8da53a29b024_MANO_GT_14.90s.mp4", "recording_id": "subject2/h2/5", "source_first_frame": "0", "source_last_frame": "446", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "d0fbe97e1d72758881a3e4f2", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/b57bc8e5cacc8da53a29b024_MANO_GT_14.90s.mp4", "bytes": 5526109, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.024356006671919138, "p90": 0.05085662134838019, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.3268972922145865, "p90": 5.337434460495999, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06503126732310832, "p90": 0.1962063169292036, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.016780621258091565, "p90": 0.04430798435847825, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06706284489948368, "p90": 0.18978106183427035, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06704700181175156, "p90": 0.18977342814953377, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "wms": { "median": 0.8213629162680249, "p90": 0.9560027108533533, "valid_pairs": 436, "valid_duration_s": 14.533333333333335, "coverage": 0.9775784753363229 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.04775194668746109, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06662976412026014, "p90": 0.22385174115132656, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01741259686550399, "p90": 0.04415179487371684, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06638594323879804, "p90": 0.21645740779579792, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06637271307297397, "p90": 0.21645013092399276, "valid_pairs": 446, "valid_duration_s": 14.866666666666667, "coverage": 1.0 }, "wms": { "median": 0.7976290161480186, "p90": 0.9579438548409775, "valid_pairs": 441, "valid_duration_s": 14.7, "coverage": 0.9887892376681614 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.37520907425022904, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05085662134838019, "camera_angular_speed_p90_deg_s": 5.337434460495999, "camera_valid_fraction": 1.0, "camera_pairs": 446, "hand_world_speed_p90_m_s": 0.22385174115132656, "hand_camera_speed_p90_m_s": 0.21645013092399276, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04430798435847825, "rhcm_p90_m_s": 0.21645013092399276, "rham_p90_deg_s": null, "wms_p90": 0.9579438548409775, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "ae1cb98346d1829b8b0584977728e6e76ee2006be3d8b5462c9115c5cf088132", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/d0fbe97e1d72758881a3e4f2/reference.npz", "window_s": [ 0.0, 14.866666666666667 ], "timestamps_source": "source frame timeline", "frame_count": 447, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e6ea711c4a051e5328de51dfbec7d1fe00bc92a17de8451ed368c146a06e2f96", "mano_sha256": "c6d355a793b13dd692519863648effadeec198d9b26d6ee2675140994b0e9b89", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "d37e5d736a31247b8468ee07", "sample_id": "cde35b090f659e100484653e", "file": "20_seconds/H2O/cde35b090f659e100484653e_MANO_GT_20.00s.mp4", "recording_id": "subject2/h1/1", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "d37e5d736a31247b8468ee07", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/cde35b090f659e100484653e_MANO_GT_20.00s.mp4", "bytes": 8122254, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0284885239823693, "p90": 0.058542771116724505, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 2.6507287244254445, "p90": 6.374736186753002, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.0627986323545286, "p90": 0.2145304621984108, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.021108460524329433, "p90": 0.04829426941084495, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.0656731311730731, "p90": 0.20889607859617343, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.06567952289435051, "p90": 0.2088896213675059, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.7800786056816043, "p90": 0.9581422525053513, "valid_pairs": 591, "valid_duration_s": 19.699999999999996, "coverage": 0.9866444073455758 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.10954455865684724, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.04106805196566077, "p90": 0.13914615468840022, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.020202622864070693, "p90": 0.056850840277481496, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.04198622100324373, "p90": 0.11718671708568568, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.04195807439963151, "p90": 0.11719697022166517, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.6753949242229345, "p90": 0.9383143853652852, "valid_pairs": 584, "valid_duration_s": 19.466666666666665, "coverage": 0.9749582637729549 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.42227212578377554, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.058542771116724505, "camera_angular_speed_p90_deg_s": 6.374736186753002, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2145304621984108, "hand_camera_speed_p90_m_s": 0.2088896213675059, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.056850840277481496, "rhcm_p90_m_s": 0.2088896213675059, "rham_p90_deg_s": null, "wms_p90": 0.9581422525053513, "gt_coverage": 1.0000000000000002, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "b4050624f7a83e65541aa52000f1a95d38b6dedfe571773fa4c0e4abd2eed988", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/d37e5d736a31247b8468ee07/reference.npz", "window_s": [ 0.03333333333333333, 20.0 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5da27fb7f7837b8d3c7bb3cdb3ce8c204de751393e540316017e4cecca260986", "mano_sha256": "fcf804a77d201f9d2d8b1843c2de5a1126c77a6b8c747a2aad3d30e4f72d4b9c", "duration_s": 14.134, "frames": 424, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "d77e4c8bc4e513a4c5b407fa", "sample_id": "7f05fb3067b5a7589e578a78", "file": "short_originals_under_20s/H2O/7f05fb3067b5a7589e578a78_MANO_GT_14.13s.mp4", "recording_id": "subject2/o2/0", "source_first_frame": "0", "source_last_frame": "423", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "d77e4c8bc4e513a4c5b407fa", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/7f05fb3067b5a7589e578a78_MANO_GT_14.13s.mp4", "bytes": 4283296, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.023100385178556306, "p90": 0.038636658985378276, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "camera_rotation": { "median": 2.754975040625699, "p90": 5.251115764576185, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06901356670511645, "p90": 0.3236926000675297, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "cmh_m_s": { "median": 0.019262603806336976, "p90": 0.04703780055129427, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06717763915628187, "p90": 0.31891112570623054, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06716847530583774, "p90": 0.3188615491088723, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "wms": { "median": 0.8270087424242706, "p90": 0.9550203459885217, "valid_pairs": 417, "valid_duration_s": 13.9, "coverage": 0.9858156028368795 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.40973365093800557, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09650701062574084, "p90": 0.284219320723764, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "cmh_m_s": { "median": 0.020994067429649255, "p90": 0.04408420129701739, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.09191286770230818, "p90": 0.2800942800591907, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.09194061648503205, "p90": 0.2800781787849635, "valid_pairs": 423, "valid_duration_s": 14.1, "coverage": 1.0 }, "wms": { "median": 0.8099758853385401, "p90": 0.9536443003948398, "valid_pairs": 419, "valid_duration_s": 13.966666666666667, "coverage": 0.9905437352245863 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.31787404640909483, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.038636658985378276, "camera_angular_speed_p90_deg_s": 5.251115764576185, "camera_valid_fraction": 1.0, "camera_pairs": 423, "hand_world_speed_p90_m_s": 0.3236926000675297, "hand_camera_speed_p90_m_s": 0.3188615491088723, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.04703780055129427, "rhcm_p90_m_s": 0.3188615491088723, "rham_p90_deg_s": null, "wms_p90": 0.9550203459885217, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "ad05a57b6059649dde716f725899d9336eeca8bcd5803ed6ae2950a70c15a368", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/d77e4c8bc4e513a4c5b407fa/reference.npz", "window_s": [ 0.0, 14.1 ], "timestamps_source": "source frame timeline", "frame_count": 424, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d82f371480db4f137786f3687f7f43dc55bcdbe7a5f46f7e9694cb61aefacbab", "mano_sha256": "70d7f07ffa16e245f54b2fef2db3eca2a93c89719544b661f4d5e6bf7b6bce99", "duration_s": 16.334, "frames": 490, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "dfb578623738b31ef8c713dc", "sample_id": "db9483a92b49248a17f49132", "file": "short_originals_under_20s/H2O/db9483a92b49248a17f49132_MANO_GT_16.33s.mp4", "recording_id": "subject2/k1/0", "source_first_frame": "0", "source_last_frame": "489", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "dfb578623738b31ef8c713dc", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/db9483a92b49248a17f49132_MANO_GT_16.33s.mp4", "bytes": 5504444, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.020187357621940207, "p90": 0.04000224174694198, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "camera_rotation": { "median": 2.1401603943410823, "p90": 5.231573035193489, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.05432067155028667, "p90": 0.34204309851819364, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "cmh_m_s": { "median": 0.015847896581194076, "p90": 0.036108799440958086, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05678595521981419, "p90": 0.3394619317689898, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.05678555627596375, "p90": 0.33948182151947615, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "wms": { "median": 0.8544738194342296, "p90": 0.9585787911037797, "valid_pairs": 419, "valid_duration_s": 13.966666666666669, "coverage": 0.8568507157464214 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4954421837124673, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08915152385477014, "p90": 0.2947703042516847, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "cmh_m_s": { "median": 0.016763834458557933, "p90": 0.03901258534757141, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08795302740919896, "p90": 0.2921946312736648, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.08800220996094318, "p90": 0.29220262517982465, "valid_pairs": 489, "valid_duration_s": 16.3, "coverage": 1.0 }, "wms": { "median": 0.8569496006317924, "p90": 0.9656751627130583, "valid_pairs": 478, "valid_duration_s": 15.933333333333334, "coverage": 0.9775051124744376 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4353756206294246, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04000224174694198, "camera_angular_speed_p90_deg_s": 5.231573035193489, "camera_valid_fraction": 1.0, "camera_pairs": 489, "hand_world_speed_p90_m_s": 0.34204309851819364, "hand_camera_speed_p90_m_s": 0.33948182151947615, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.03901258534757141, "rhcm_p90_m_s": 0.33948182151947615, "rham_p90_deg_s": null, "wms_p90": 0.9656751627130583, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "82764115b5462c4d83d6a6b66cc8411f9711f52933c6d69ad686a06becd462ce", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/dfb578623738b31ef8c713dc/reference.npz", "window_s": [ 0.0, 16.3 ], "timestamps_source": "source frame timeline", "frame_count": 490, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c4dd202748f9ecf0b0c5add4fdb8686ab3459190697f7c4c858920c32384ccd9", "mano_sha256": "fa8f1cb40ffa1b0e3e29f0f7c6999a188fa4ff90f58290d8c9221388b3a1a278", "duration_s": 18.534, "frames": 556, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e008a6173b1c484c35f095bb", "sample_id": "c652a372065f11cf5be679d2", "file": "short_originals_under_20s/H2O/c652a372065f11cf5be679d2_MANO_GT_18.53s.mp4", "recording_id": "subject1/k1/4", "source_first_frame": "0", "source_last_frame": "555", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "e008a6173b1c484c35f095bb", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/c652a372065f11cf5be679d2_MANO_GT_18.53s.mp4", "bytes": 6273731, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02884957970053432, "p90": 0.08455127844211906, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 1.0 }, "camera_rotation": { "median": 3.5154206240435033, "p90": 10.395458026705477, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03046577195132719, "p90": 0.10553111352938038, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 1.0 }, "cmh_m_s": { "median": 0.023035677745927543, "p90": 0.05860429864137085, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.0452123908347879, "p90": 0.1154406915192518, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.045193239135996176, "p90": 0.1154387825753459, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 1.0 }, "wms": { "median": 0.7821696067257735, "p90": 0.9532207964619639, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.9711711711711712 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.11631217223508732, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07071566964929757, "p90": 0.39126796414770815, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.9441441441441443 }, "cmh_m_s": { "median": 0.02361136084465818, "p90": 0.06198960227457629, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.9441441441441443 }, "rhcm_continuous_estimate_m_s": { "median": 0.07325446899988369, "p90": 0.377584967345212, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.9441441441441443 }, "rhcm_discrete_m_s": { "median": 0.07325729173874714, "p90": 0.376755147507287, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.9441441441441443 }, "wms": { "median": 0.8204466091230612, "p90": 0.9532630004571541, "valid_pairs": 520, "valid_duration_s": 17.333333333333336, "coverage": 0.936936936936937 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.25930211367602807, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.08455127844211906, "camera_angular_speed_p90_deg_s": 10.395458026705477, "camera_valid_fraction": 1.0, "camera_pairs": 555, "hand_world_speed_p90_m_s": 0.39126796414770815, "hand_camera_speed_p90_m_s": 0.376755147507287, "hand_world_valid_fraction": 0.9730215827338129, "hand_camera_valid_fraction": 0.9730215827338129, "cmh_p90_m_s": 0.06198960227457629, "rhcm_p90_m_s": 0.376755147507287, "rham_p90_deg_s": null, "wms_p90": 0.9532630004571541, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "cf13a3b064e21d8b1e27df9004f0d7068a50175a2e8c78b42ad325918149ea6e", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/e008a6173b1c484c35f095bb/reference.npz", "window_s": [ 0.0, 18.5 ], "timestamps_source": "source frame timeline", "frame_count": 556, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "fb880d5774b7a6728224c4e41d691229c0d54d26f47491a7a4db3b114d8a3133", "mano_sha256": "3d1c2a1f5746f64c66cf6602c91b95906964ebd89d30d9a7738df25a1b040d93", "duration_s": 11.7, "frames": 351, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e4725b1311d0054c285e24c0", "sample_id": "950552c9e96d4cf44491b4fa", "file": "short_originals_under_20s/H2O/950552c9e96d4cf44491b4fa_MANO_GT_11.70s.mp4", "recording_id": "subject1/k1/0", "source_first_frame": "0", "source_last_frame": "350", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "e4725b1311d0054c285e24c0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/950552c9e96d4cf44491b4fa_MANO_GT_11.70s.mp4", "bytes": 3988438, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03302174209682898, "p90": 0.08604103568842678, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "camera_rotation": { "median": 3.7101970220225886, "p90": 11.255151349156256, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.042283911145960903, "p90": 0.11246430025965651, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.021179969318981633, "p90": 0.04828035257345355, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.053460725502740704, "p90": 0.11787204838162822, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0534269974724534, "p90": 0.11784300013021673, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "wms": { "median": 0.7970878414120314, "p90": 0.9526818277938172, "valid_pairs": 341, "valid_duration_s": 11.366666666666664, "coverage": 0.9742857142857141 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": -0.02573170859703811, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09369786123037059, "p90": 0.25539163180782015, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.023830896162360896, "p90": 0.06752200202437648, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.09840139136211834, "p90": 0.24607878378281597, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.09826056757767163, "p90": 0.24637744141242998, "valid_pairs": 350, "valid_duration_s": 11.666666666666666, "coverage": 1.0 }, "wms": { "median": 0.8281041848401821, "p90": 0.9466506355140676, "valid_pairs": 346, "valid_duration_s": 11.533333333333331, "coverage": 0.9885714285714284 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.08348667569297942, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.08604103568842678, "camera_angular_speed_p90_deg_s": 11.255151349156256, "camera_valid_fraction": 1.0, "camera_pairs": 350, "hand_world_speed_p90_m_s": 0.25539163180782015, "hand_camera_speed_p90_m_s": 0.24637744141242998, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.06752200202437648, "rhcm_p90_m_s": 0.24637744141242998, "rham_p90_deg_s": null, "wms_p90": 0.9526818277938172, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "5671b4a79cbb89532227ae59762cff7ca4dda564606282a6964efd660bb79b6e", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/e4725b1311d0054c285e24c0/reference.npz", "window_s": [ 0.0, 11.666666666666666 ], "timestamps_source": "source frame timeline", "frame_count": 351, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ad8c829f97aef3648ed6e75312b553cd7a53a947b369f304a02484b1b2a97d8c", "mano_sha256": "f001ad37eb38161582ba4fdefcb9a07cca8d9175306f7afe5c06b0bf0ca2ddfb", "duration_s": 12.167, "frames": 365, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e9ffe788c419504ccf8ee977", "sample_id": "198d8faa7cc3d046c477325c", "file": "short_originals_under_20s/H2O/198d8faa7cc3d046c477325c_MANO_GT_12.17s.mp4", "recording_id": "subject1/o2/7", "source_first_frame": "0", "source_last_frame": "364", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "e9ffe788c419504ccf8ee977", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/198d8faa7cc3d046c477325c_MANO_GT_12.17s.mp4", "bytes": 3709322, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.026906822751248304, "p90": 0.05571508633402548, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 2.499204800987531, "p90": 5.392224751262293, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.048598909496448364, "p90": 0.13233767527260307, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "cmh_m_s": { "median": 0.018755611136557652, "p90": 0.04490129028343559, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.04931327829166385, "p90": 0.12343506816531145, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.04931259642194889, "p90": 0.12336819464608798, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "wms": { "median": 0.7279360357039905, "p90": 0.936978178250456, "valid_pairs": 353, "valid_duration_s": 11.766666666666666, "coverage": 0.9697802197802198 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.36417281348788194, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.056569706298602584, "p90": 0.20726861836846885, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "cmh_m_s": { "median": 0.019374220570268365, "p90": 0.045204547613553796, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05548231461990355, "p90": 0.20119822910890556, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.05547285995524397, "p90": 0.20121092407677316, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 1.0 }, "wms": { "median": 0.7793110231994222, "p90": 0.96171187947191, "valid_pairs": 362, "valid_duration_s": 12.066666666666666, "coverage": 0.9945054945054945 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.16872218690400506, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05571508633402548, "camera_angular_speed_p90_deg_s": 5.392224751262293, "camera_valid_fraction": 1.0, "camera_pairs": 364, "hand_world_speed_p90_m_s": 0.20726861836846885, "hand_camera_speed_p90_m_s": 0.20121092407677316, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.045204547613553796, "rhcm_p90_m_s": 0.20121092407677316, "rham_p90_deg_s": null, "wms_p90": 0.96171187947191, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "2fbc4b97471c6f4fada170e437d1b6767d0c02222e5b524fd2baf93707e88ca9", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/e9ffe788c419504ccf8ee977/reference.npz", "window_s": [ 0.0, 12.133333333333333 ], "timestamps_source": "source frame timeline", "frame_count": 365, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c026e963aa8aec15ffe722033628e1268e088da0627ea2731237640783c42d40", "mano_sha256": "e497d10c623baadfdba0bfd06732e3fb1442e88d6c925f479c049d79fe116f67", "duration_s": 13.6, "frames": 408, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "f0f922c3daa41c15a71b4007", "sample_id": "e2047d2acdb1c54fe3d347e8", "file": "short_originals_under_20s/H2O/e2047d2acdb1c54fe3d347e8_MANO_GT_13.60s.mp4", "recording_id": "subject3/h1/0", "source_first_frame": "0", "source_last_frame": "407", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "f0f922c3daa41c15a71b4007", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/e2047d2acdb1c54fe3d347e8_MANO_GT_13.60s.mp4", "bytes": 4348689, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.017119532809959124, "p90": 0.03474248655698694, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "camera_rotation": { "median": 1.4663377878845334, "p90": 3.155845292850888, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03049712393214728, "p90": 0.1739614758699268, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.010717087519883725, "p90": 0.02228184721292915, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.03279681677885539, "p90": 0.17164682075118193, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.03278484025933916, "p90": 0.17165666784006425, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "wms": { "median": 0.8205441752465377, "p90": 0.9574155532347077, "valid_pairs": 340, "valid_duration_s": 11.333333333333332, "coverage": 0.8353808353808353 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.40589488358251036, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.04618343813389771, "p90": 0.1204441636231561, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "cmh_m_s": { "median": 0.011436182518514611, "p90": 0.024131512763408008, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.048647584237411966, "p90": 0.12343038018175173, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.04864754419766479, "p90": 0.12343191282398253, "valid_pairs": 407, "valid_duration_s": 13.566666666666666, "coverage": 1.0 }, "wms": { "median": 0.8385216020263596, "p90": 0.9552333528730839, "valid_pairs": 396, "valid_duration_s": 13.200000000000001, "coverage": 0.972972972972973 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.19850369191139983, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.03474248655698694, "camera_angular_speed_p90_deg_s": 3.155845292850888, "camera_valid_fraction": 1.0, "camera_pairs": 407, "hand_world_speed_p90_m_s": 0.1739614758699268, "hand_camera_speed_p90_m_s": 0.17165666784006425, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.024131512763408008, "rhcm_p90_m_s": 0.17165666784006425, "rham_p90_deg_s": null, "wms_p90": 0.9574155532347077, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "66c711015664244e3b5975baff35043fa18168874700b5912c1f79156b46789c", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/f0f922c3daa41c15a71b4007/reference.npz", "window_s": [ 0.0, 13.566666666666666 ], "timestamps_source": "source frame timeline", "frame_count": 408, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "1a58f0b760a13f59d6984df657cd45e46c7bdee66d6a89b2f8d698aee476a150", "mano_sha256": "845ed7847ae78dc1771a24571eb93d139ab4a2b715d4519bb7ae82c3c8cdd38e", "duration_s": 19.2, "frames": 576, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "f478ac3ea350992fb2bf7f61", "sample_id": "df45adfd9538a7e4745efea2", "file": "short_originals_under_20s/H2O/df45adfd9538a7e4745efea2_MANO_GT_19.20s.mp4", "recording_id": "subject2/h1/6", "source_first_frame": "0", "source_last_frame": "575", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "f478ac3ea350992fb2bf7f61", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/H2O/df45adfd9538a7e4745efea2_MANO_GT_19.20s.mp4", "bytes": 7734055, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02337897078079833, "p90": 0.046749708013357795, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "camera_rotation": { "median": 2.1720580130095604, "p90": 4.3742879245745385, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06251833644541625, "p90": 0.14422343476722896, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "cmh_m_s": { "median": 0.01578804685445272, "p90": 0.03298493569549808, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.06142612605107085, "p90": 0.13792704001044473, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.06144355128939108, "p90": 0.13792884989728676, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "wms": { "median": 0.8005113768655356, "p90": 0.9467077387646841, "valid_pairs": 569, "valid_duration_s": 18.96666666666667, "coverage": 0.9895652173913044 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.2235298944604353, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.04541277126571449, "p90": 0.10940186361256705, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "cmh_m_s": { "median": 0.014308532108686526, "p90": 0.030799664151398187, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.04668708235830152, "p90": 0.11103612358951669, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.046692635383110484, "p90": 0.11103991403664912, "valid_pairs": 575, "valid_duration_s": 19.166666666666668, "coverage": 1.0 }, "wms": { "median": 0.8102909103189462, "p90": 0.9560927303321749, "valid_pairs": 556, "valid_duration_s": 18.533333333333335, "coverage": 0.9669565217391305 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.20530778164924507, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.046749708013357795, "camera_angular_speed_p90_deg_s": 4.3742879245745385, "camera_valid_fraction": 1.0, "camera_pairs": 575, "hand_world_speed_p90_m_s": 0.14422343476722896, "hand_camera_speed_p90_m_s": 0.13792884989728676, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.03298493569549808, "rhcm_p90_m_s": 0.13792884989728676, "rham_p90_deg_s": null, "wms_p90": 0.9560927303321749, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "6d1e82e8605b11a9004fa65c803a791a4e99aa87d17296a268386fc916b4f1fb", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/f478ac3ea350992fb2bf7f61/reference.npz", "window_s": [ 0.0, 19.166666666666668 ], "timestamps_source": "source frame timeline", "frame_count": 576, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "H2O", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "cfe567b40d465bc255286db78d285ecc64d61e940d26041fcc5c032a3cd7f8c0", "mano_sha256": "16db53090c11357c413d44c29fed425d67ac0974e198580fe226b4bab63e4342", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native H2O hand_pose_mano; official fitted annotation", "scene_depth_source": "native aligned sensor depth, millimetres converted to metres", "scene_camera_source": "native calibrated camera pose; stationary when the source camera is stationary", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "fb7c82e6427650bac0c3f633", "sample_id": "fb157c309c02dc5b6cf3c4cc", "file": "20_seconds/H2O/fb157c309c02dc5b6cf3c4cc_MANO_GT_20.00s.mp4", "recording_id": "subject1/h2/7", "source_first_frame": "86", "source_last_frame": "685", "source_split": "official_train", "training_label_source": "H2O source 21-joint hand_pose annotations", "quality_warnings": "", "id": "fb7c82e6427650bac0c3f633", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/H2O/fb157c309c02dc5b6cf3c4cc_MANO_GT_20.00s.mp4", "bytes": 6831521, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.020008741791636964, "p90": 0.04000274603972476, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 2.1317183715937373, "p90": 4.590662699369954, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.02284059106660948, "p90": 0.06889600891251037, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.01203705749998181, "p90": 0.02760155131599165, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.025922157198574245, "p90": 0.07153096561282062, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.02592180977865828, "p90": 0.07153373644761994, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.751980312190498, "p90": 0.9382673858184065, "valid_pairs": 522, "valid_duration_s": 17.400000000000006, "coverage": 0.871452420701169 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.21587629326469418, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.05812557754263361, "p90": 0.23950697686770922, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "cmh_m_s": { "median": 0.013800468216294425, "p90": 0.02901692427426277, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_continuous_estimate_m_s": { "median": 0.06002815675852362, "p90": 0.2396597588763239, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "rhcm_discrete_m_s": { "median": 0.06019307597411493, "p90": 0.23966914074454918, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "wms": { "median": 0.849711189282561, "p90": 0.9682757475516564, "valid_pairs": 591, "valid_duration_s": 19.7, "coverage": 0.986644407345576 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.10812441024896564, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04000274603972476, "camera_angular_speed_p90_deg_s": 4.590662699369954, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.23950697686770922, "hand_camera_speed_p90_m_s": 0.23966914074454918, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.02901692427426277, "rhcm_p90_m_s": 0.23966914074454918, "rham_p90_deg_s": null, "wms_p90": 0.9682757475516564, "gt_coverage": 1.0000000000000002, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "H2O published camera pose", "hand_source": "H2O source 21-joint hand_pose annotations", "reference_sha256": "23ac2aa51929924992e9cf5e95a751799aef43dedb2ba5da46e5c42ddd9ea544", "metadata_sha256": "41cda975bef2aaebb8611ebdbde36041671cef9dc0c31957142106853ded5dd4", "source_file": "long20_20260918/local/train/previews/fb7c82e6427650bac0c3f633/reference.npz", "window_s": [ 2.8666666666666667, 22.833333333333332 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native calibrated camera pose; stationary when the source camera is stationary", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方位姿/标定", "kind": "official", "detail": "使用数据集发布的相机位姿与标定。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "33fe9463529dc8e44b3116679cc7c56c099cc71ce0300263a0a685d3ee51d0ea", "mano_sha256": "83ee3ad16bc45716962524298d250087aa3dad4804f47b22620a2b846153a7e0", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "03d0921a7bf4b1006e591b1b", "sample_id": "a7e05e5ec3ee2a9ceee1384a", "file": "20_seconds/HOT3D/a7e05e5ec3ee2a9ceee1384a_MANO_GT_20.00s.mp4", "recording_id": "P0012_119de519", "source_first_frame": "987", "source_last_frame": "1586", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 55/600 output frames; may include genuine out-of-view frames; per-hand official MANO coverage: left absent 163/600, right absent 55/600; absence alone does not establish visible-hand label failure", "id": "03d0921a7bf4b1006e591b1b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/a7e05e5ec3ee2a9ceee1384a_MANO_GT_20.00s.mp4", "bytes": 4895861, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.04984736127051999, "p90": 0.2957904775763267, "valid_pairs": 599, "valid_duration_s": 19.966647078996175, "coverage": 1.0 }, "camera_rotation": { "median": 9.578523896846223, "p90": 28.655699872939266, "valid_pairs": 599, "valid_duration_s": 19.966647078996175, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07692015119017803, "p90": 0.398178999435072, "valid_pairs": 436, "valid_duration_s": 14.533291624000412, "coverage": 0.7278784247801247 }, "cmh_m_s": { "median": 0.07506075203194157, "p90": 0.30748448344468565, "valid_pairs": 436, "valid_duration_s": 14.533291624000412, "coverage": 0.7278784247801247 }, "rhcm_continuous_estimate_m_s": { "median": 0.09101454189041953, "p90": 0.26308795271923935, "valid_pairs": 436, "valid_duration_s": 14.533291624000412, "coverage": 0.7278784247801247 }, "rhcm_discrete_m_s": { "median": 0.09101861187051234, "p90": 0.2631294242792725, "valid_pairs": 436, "valid_duration_s": 14.533291624000412, "coverage": 0.7278784247801247 }, "wms": { "median": 0.5856800288213433, "p90": 0.9177600992871504, "valid_pairs": 433, "valid_duration_s": 14.433309157000622, "coverage": 0.7228709507358236 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5936928686791323, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.11193291695206656, "p90": 0.3790664884128411, "valid_pairs": 543, "valid_duration_s": 18.099895668994577, "coverage": 0.9065065154597078 }, "cmh_m_s": { "median": 0.09141149216199287, "p90": 0.3223764153845734, "valid_pairs": 543, "valid_duration_s": 18.099895668994577, "coverage": 0.9065065154597078 }, "rhcm_continuous_estimate_m_s": { "median": 0.10164486294422166, "p90": 0.3076799268895393, "valid_pairs": 543, "valid_duration_s": 18.099895668994577, "coverage": 0.9065065154597078 }, "rhcm_discrete_m_s": { "median": 0.1015916035293688, "p90": 0.3082496817277767, "valid_pairs": 543, "valid_duration_s": 18.099895668994577, "coverage": 0.9065065154597078 }, "wms": { "median": 0.5821078516952246, "p90": 0.9257087920751369, "valid_pairs": 542, "valid_duration_s": 18.06487019499764, "coverage": 0.9047523163766888 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6330660345114989, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.2957904775763267, "camera_angular_speed_p90_deg_s": 28.655699872939266, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.398178999435072, "hand_camera_speed_p90_m_s": 0.3082496817277767, "hand_world_valid_fraction": 0.8183333333333334, "hand_camera_valid_fraction": 0.8183333333333334, "cmh_p90_m_s": 0.3223764153845734, "rhcm_p90_m_s": 0.3082496817277767, "rham_p90_deg_s": null, "wms_p90": 0.9257087920751369, "gt_coverage": 0.9065065154597078, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "fc57f721370d50ff6c80d35b88dc5e8013bbae29005b5fb38398c6709c551769", "metadata_sha256": "24467b60495ba60ab27d263ee559737d1f1e095d942b3f0c0c8b113683b257c6", "source_file": "long20_20260918/local/train/previews/03d0921a7bf4b1006e591b1b/dense_reference.npz", "window_s": [ 45229.908223685, 45249.874870764 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "1f772bec611d273fad6c367de10eb20d2823abf02ed82199daec43dec6c31cb2", "mano_sha256": "10f7e2b24ec034b7b9ef5c72d9df105274c6c66fb275c0184940c1b523f2439d", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "0981b2b39383c9d2f884d3f4", "sample_id": "e5e0bee04850db4cde30baac", "file": "20_seconds/HOT3D/e5e0bee04850db4cde30baac_MANO_GT_20.00s.mp4", "recording_id": "P0012_0a21e4c2", "source_first_frame": "3068", "source_last_frame": "3667", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "0981b2b39383c9d2f884d3f4", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/e5e0bee04850db4cde30baac_MANO_GT_20.00s.mp4", "bytes": 6119589, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.10932615238712473, "p90": 0.23914605828777918, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "camera_rotation": { "median": 17.569694241527767, "p90": 46.86320323523543, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.09741345450935386, "p90": 0.5705334150940944, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "cmh_m_s": { "median": 0.12461591003407899, "p90": 0.32642331875468206, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.16930086841404746, "p90": 0.5282257751425046, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.16918947058270958, "p90": 0.5281938542595691, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "wms": { "median": 0.6590040968968477, "p90": 0.8882774512150808, "valid_pairs": 576, "valid_duration_s": 19.200008080006228, "coverage": 0.9616021786456083 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5006631453760727, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.12252904050096036, "p90": 0.4523224295145304, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "cmh_m_s": { "median": 0.12211552383282717, "p90": 0.345098698944254, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.16628573176194114, "p90": 0.5185340105953206, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.16703787980306956, "p90": 0.5189838547470679, "valid_pairs": 599, "valid_duration_s": 19.966685295003117, "coverage": 1.0 }, "wms": { "median": 0.6373784945860594, "p90": 0.9210905196659751, "valid_pairs": 597, "valid_duration_s": 19.899996992011438, "coverage": 0.996660021330212 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4846386675674619, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.23914605828777918, "camera_angular_speed_p90_deg_s": 46.86320323523543, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5705334150940944, "hand_camera_speed_p90_m_s": 0.5281938542595691, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.345098698944254, "rhcm_p90_m_s": 0.5281938542595691, "rham_p90_deg_s": null, "wms_p90": 0.9210905196659751, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "0b954aa3a65874dc8d623ab39cfa64d2393a2e74d5ca20c4bb20d052a109780a", "metadata_sha256": "e384bfa5a1cfa45059be468cd8609a1508af134193cefaa46cf394c66c2bf2af", "source_file": "long20_20260918/local/train/previews/0981b2b39383c9d2f884d3f4/dense_reference.npz", "window_s": [ 47167.274853748, 47187.241539043 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7582fa0fa28782755244f6066ae3b67e370de92a6de9d63f7bdfb0a272c10023", "mano_sha256": "e83bc03ba5ba4ab2e6bc9c2267f09b4c23f362b2067fed71bbb39af548f774cf", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "0acec37856d8623f0686ac2d", "sample_id": "27e1b004683cacb565424e09", "file": "20_seconds/HOT3D/27e1b004683cacb565424e09_MANO_GT_20.00s.mp4", "recording_id": "P0009_b1e71d7f", "source_first_frame": "75", "source_last_frame": "674", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 68/600, right absent 0/600; absence alone does not establish visible-hand label failure", "id": "0acec37856d8623f0686ac2d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/27e1b004683cacb565424e09_MANO_GT_20.00s.mp4", "bytes": 7284083, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05998138500891533, "p90": 0.3039747024979961, "valid_pairs": 599, "valid_duration_s": 19.96666319799988, "coverage": 1.0 }, "camera_rotation": { "median": 11.748354845502908, "p90": 31.605384032596042, "valid_pairs": 599, "valid_duration_s": 19.96666319799988, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.09576458123583181, "p90": 0.29178364029628473, "valid_pairs": 530, "valid_duration_s": 17.666641877003713, "coverage": 0.8848069255143961 }, "cmh_m_s": { "median": 0.07839669099823826, "p90": 0.2969959187786372, "valid_pairs": 530, "valid_duration_s": 17.666641877003713, "coverage": 0.8848069255143961 }, "rhcm_continuous_estimate_m_s": { "median": 0.09674711924589541, "p90": 0.30036449023410283, "valid_pairs": 530, "valid_duration_s": 17.666641877003713, "coverage": 0.8848069255143961 }, "rhcm_discrete_m_s": { "median": 0.09643609539687731, "p90": 0.3000010455084423, "valid_pairs": 530, "valid_duration_s": 17.666641877003713, "coverage": 0.8848069255143961 }, "wms": { "median": 0.5471001381309586, "p90": 0.8555279757457523, "valid_pairs": 530, "valid_duration_s": 17.666641877003713, "coverage": 0.8848069255143961 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.7085359963196335, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1266054299721912, "p90": 0.5314945325391852, "valid_pairs": 599, "valid_duration_s": 19.96666319799988, "coverage": 1.0 }, "cmh_m_s": { "median": 0.07980252323386414, "p90": 0.32249321278030396, "valid_pairs": 599, "valid_duration_s": 19.96666319799988, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.14647226381109224, "p90": 0.3745848360537108, "valid_pairs": 599, "valid_duration_s": 19.96666319799988, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.1463426244437607, "p90": 0.3745971407705144, "valid_pairs": 599, "valid_duration_s": 19.96666319799988, "coverage": 1.0 }, "wms": { "median": 0.566086756338437, "p90": 0.8830452064350193, "valid_pairs": 599, "valid_duration_s": 19.96666319799988, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.704302153533481, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.3039747024979961, "camera_angular_speed_p90_deg_s": 31.605384032596042, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5314945325391852, "hand_camera_speed_p90_m_s": 0.3745971407705144, "hand_world_valid_fraction": 0.9433333333333334, "hand_camera_valid_fraction": 0.9433333333333334, "cmh_p90_m_s": 0.32249321278030396, "rhcm_p90_m_s": 0.3745971407705144, "rham_p90_deg_s": null, "wms_p90": 0.8830452064350193, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "e97146e553986077c5c8b06432d2a778062a2a047a0ab4ea4adeda2c19be4405", "metadata_sha256": "1e4edbe0166f16eced80cb8b4f3bc6e08d86ea2d542f5a566a5ba808a36f62b8", "source_file": "long20_20260918/local/benchmark/previews/0acec37856d8623f0686ac2d/dense_reference.npz", "window_s": [ 38926.508193533, 38946.474856731 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c09c0930a5064842f801cedd1eab8316b1c49688d896231cf07a44d4114d82c7", "mano_sha256": "9eba273ae19fdaeb4ec867bc842245bb1087ba3cb50e205627415600803bb198", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "0b0d0b1190158d31e39b3ce4", "sample_id": "50b76925f3802854468ed4bd", "file": "20_seconds/HOT3D/50b76925f3802854468ed4bd_MANO_GT_20.00s.mp4", "recording_id": "P0015_e3f96b3f", "source_first_frame": "1776", "source_last_frame": "2375", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 0/600, right absent 207/600; absence alone does not establish visible-hand label failure", "id": "0b0d0b1190158d31e39b3ce4", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/50b76925f3802854468ed4bd_MANO_GT_20.00s.mp4", "bytes": 5007507, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.027142337125480698, "p90": 0.16804196065336927, "valid_pairs": 599, "valid_duration_s": 19.9666871369991, "coverage": 1.0 }, "camera_rotation": { "median": 5.77860009683198, "p90": 38.64367293496782, "valid_pairs": 599, "valid_duration_s": 19.9666871369991, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.02514195973987062, "p90": 0.4192922028556987, "valid_pairs": 599, "valid_duration_s": 19.9666871369991, "coverage": 1.0 }, "cmh_m_s": { "median": 0.036624916006054706, "p90": 0.2760137014951324, "valid_pairs": 599, "valid_duration_s": 19.9666871369991, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.042701779146771195, "p90": 0.434961195429446, "valid_pairs": 599, "valid_duration_s": 19.9666871369991, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.04265328324733738, "p90": 0.4370364929007014, "valid_pairs": 599, "valid_duration_s": 19.9666871369991, "coverage": 1.0 }, "wms": { "median": 0.70282253626168, "p90": 0.9438607471379648, "valid_pairs": 410, "valid_duration_s": 13.666718567023054, "coverage": 0.6844760211471466 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.8201223890430538, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06747946932677788, "p90": 0.3984815305471295, "valid_pairs": 391, "valid_duration_s": 13.033350274003169, "coverage": 0.652754770211821 }, "cmh_m_s": { "median": 0.10510468674325962, "p90": 0.3358938582235557, "valid_pairs": 391, "valid_duration_s": 13.033350274003169, "coverage": 0.652754770211821 }, "rhcm_continuous_estimate_m_s": { "median": 0.1336025674245611, "p90": 0.46355055213077184, "valid_pairs": 391, "valid_duration_s": 13.033350274003169, "coverage": 0.652754770211821 }, "rhcm_discrete_m_s": { "median": 0.13360097436180693, "p90": 0.46396081743570866, "valid_pairs": 391, "valid_duration_s": 13.033350274003169, "coverage": 0.652754770211821 }, "wms": { "median": 0.7287917427731394, "p90": 0.9592612700381926, "valid_pairs": 380, "valid_duration_s": 12.66666202199849, "coverage": 0.6343897680715717 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5469160908510562, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.16804196065336927, "camera_angular_speed_p90_deg_s": 38.64367293496782, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4192922028556987, "hand_camera_speed_p90_m_s": 0.46396081743570866, "hand_world_valid_fraction": 0.8275, "hand_camera_valid_fraction": 0.8275, "cmh_p90_m_s": 0.3358938582235557, "rhcm_p90_m_s": 0.46396081743570866, "rham_p90_deg_s": null, "wms_p90": 0.9592612700381926, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "e19089b4892e50d42009c03f753cb1877b664d5c053e813bb4ea54f0024cbdf1", "metadata_sha256": "c7eab2b5ebe8bfe5cd2f4ac795038d51269e00e45969dd39dc35565e97f0ba54", "source_file": "long20_20260918/local/train/previews/0b0d0b1190158d31e39b3ce4/dense_reference.npz", "window_s": [ 58889.208204281, 58909.174891418 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "69a5ca3aae547a9ed4b8f6ffcec42cd4a8d015340d6dd26bef8192dbb788ff30", "mano_sha256": "2f079bcd2e69c6978316846f514361d018d79c5b57746edfc334e3cb6d418b73", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "238b097ded134dc533b7a1c9", "sample_id": "ebd1d1b1ae6fe5ff6417dae3", "file": "20_seconds/HOT3D/ebd1d1b1ae6fe5ff6417dae3_MANO_GT_20.00s.mp4", "recording_id": "P0009_5f2583e0", "source_first_frame": "2475", "source_last_frame": "3074", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 45/600, right absent 0/600; absence alone does not establish visible-hand label failure", "id": "238b097ded134dc533b7a1c9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/ebd1d1b1ae6fe5ff6417dae3_MANO_GT_20.00s.mp4", "bytes": 6836851, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.14681593798558035, "p90": 0.4380254624111761, "valid_pairs": 599, "valid_duration_s": 19.966637516998162, "coverage": 1.0 }, "camera_rotation": { "median": 19.01741300413107, "p90": 58.36176832798198, "valid_pairs": 599, "valid_duration_s": 19.966637516998162, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.23225314670407265, "p90": 0.7467922073279318, "valid_pairs": 552, "valid_duration_s": 18.39989023599628, "coverage": 0.9215317411523065 }, "cmh_m_s": { "median": 0.19427664652580368, "p90": 0.5295304541520907, "valid_pairs": 552, "valid_duration_s": 18.39989023599628, "coverage": 0.9215317411523065 }, "rhcm_continuous_estimate_m_s": { "median": 0.25403384039246024, "p90": 0.8195699837936271, "valid_pairs": 552, "valid_duration_s": 18.39989023599628, "coverage": 0.9215317411523065 }, "rhcm_discrete_m_s": { "median": 0.2536009155763682, "p90": 0.812590524473569, "valid_pairs": 552, "valid_duration_s": 18.39989023599628, "coverage": 0.9215317411523065 }, "wms": { "median": 0.6388585686446305, "p90": 0.9132977548681219, "valid_pairs": 552, "valid_duration_s": 18.39989023599628, "coverage": 0.9215317411523065 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3968739195644397, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.15068889439384034, "p90": 0.5217295904966319, "valid_pairs": 599, "valid_duration_s": 19.966637516998162, "coverage": 1.0 }, "cmh_m_s": { "median": 0.1792718572094241, "p90": 0.4840376716557562, "valid_pairs": 599, "valid_duration_s": 19.966637516998162, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.19708866904402836, "p90": 0.5373181688823477, "valid_pairs": 599, "valid_duration_s": 19.966637516998162, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.19696046493372601, "p90": 0.537683171910224, "valid_pairs": 599, "valid_duration_s": 19.966637516998162, "coverage": 1.0 }, "wms": { "median": 0.5647645170778501, "p90": 0.8643542010953384, "valid_pairs": 599, "valid_duration_s": 19.966637516998162, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4824869766221293, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.4380254624111761, "camera_angular_speed_p90_deg_s": 58.36176832798198, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.7467922073279318, "hand_camera_speed_p90_m_s": 0.812590524473569, "hand_world_valid_fraction": 0.9625, "hand_camera_valid_fraction": 0.9625, "cmh_p90_m_s": 0.5295304541520907, "rhcm_p90_m_s": 0.812590524473569, "rham_p90_deg_s": null, "wms_p90": 0.9132977548681219, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "fefacb1f4f4eda4644c49faa24e701e6176ac5947ab6e9f34edc9707aa45a3b8", "metadata_sha256": "66aab89ef56516ea690d76bdd798b0e4404c29a8482830552dad5966a69f640e", "source_file": "long20_20260918/local/benchmark/previews/238b097ded134dc533b7a1c9/dense_reference.npz", "window_s": [ 41990.541548531, 42010.508186048 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "bb95aa4cf4a3c33c0f42157f727077fad4c4d6e2907867f9b5341aa5d3897ec3", "mano_sha256": "f54d43d1289abbefd467c73fa9ad108a7f0e082b9bd38d1ccecb7154e43310cb", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "27398d3a5626c6e791037278", "sample_id": "f4e2edd7d8d2c2f686129b05", "file": "20_seconds/HOT3D/f4e2edd7d8d2c2f686129b05_MANO_GT_20.00s.mp4", "recording_id": "P0009_8c0caf6c", "source_first_frame": "2775", "source_last_frame": "3374", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "", "id": "27398d3a5626c6e791037278", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/f4e2edd7d8d2c2f686129b05_MANO_GT_20.00s.mp4", "bytes": 5577290, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0885418687504954, "p90": 0.24450366360882556, "valid_pairs": 599, "valid_duration_s": 19.966703559999587, "coverage": 1.0 }, "camera_rotation": { "median": 14.172534398145114, "p90": 54.907716544425774, "valid_pairs": 599, "valid_duration_s": 19.966703559999587, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.20083692380942272, "p90": 0.5983806758766342, "valid_pairs": 599, "valid_duration_s": 19.966703559999587, "coverage": 1.0 }, "cmh_m_s": { "median": 0.11166232758287103, "p90": 0.37965255387241276, "valid_pairs": 599, "valid_duration_s": 19.966703559999587, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.20545689384897117, "p90": 0.49038007571738346, "valid_pairs": 599, "valid_duration_s": 19.966703559999587, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.20545186321277148, "p90": 0.49069802293628345, "valid_pairs": 599, "valid_duration_s": 19.966703559999587, "coverage": 1.0 }, "wms": { "median": 0.6611767099726498, "p90": 0.91452691631842, "valid_pairs": 595, "valid_duration_s": 19.833419612004946, "coverage": 0.9933246893963179 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.49821743038843996, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.18311665178759517, "p90": 0.42438229157730967, "valid_pairs": 579, "valid_duration_s": 19.300046203999955, "coverage": 0.966611546367865 }, "cmh_m_s": { "median": 0.09677137805468068, "p90": 0.2926569744217701, "valid_pairs": 579, "valid_duration_s": 19.300046203999955, "coverage": 0.966611546367865 }, "rhcm_continuous_estimate_m_s": { "median": 0.1804750890779671, "p90": 0.585869948836925, "valid_pairs": 579, "valid_duration_s": 19.300046203999955, "coverage": 0.966611546367865 }, "rhcm_discrete_m_s": { "median": 0.18047667523884126, "p90": 0.5859317951206148, "valid_pairs": 579, "valid_duration_s": 19.300046203999955, "coverage": 0.966611546367865 }, "wms": { "median": 0.7104938481956516, "p90": 0.9176752176601649, "valid_pairs": 579, "valid_duration_s": 19.300046203999955, "coverage": 0.966611546367865 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.23951492797983104, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.24450366360882556, "camera_angular_speed_p90_deg_s": 54.907716544425774, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5983806758766342, "hand_camera_speed_p90_m_s": 0.5859317951206148, "hand_world_valid_fraction": 0.9841666666666666, "hand_camera_valid_fraction": 0.9841666666666666, "cmh_p90_m_s": 0.37965255387241276, "rhcm_p90_m_s": 0.5859317951206148, "rham_p90_deg_s": null, "wms_p90": 0.9176752176601649, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "736223732c1317332e0165a73e28f42cd45ab7450f31b215b7bfc121de927b9a", "metadata_sha256": "48b66ac622c3a8c332c3676d92debc730d3c6a03d3f2c1924c822024544eaafc", "source_file": "long20_20260918/local/benchmark/previews/27398d3a5626c6e791037278/dense_reference.npz", "window_s": [ 43451.50817269, 43471.47487625 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b4744210806c6cc647af0c9c3c71923947f28c0505ce7ff214310ebc91fe7804", "mano_sha256": "e90ddf92d3190b60b85a7d7db8cd4a50485ca7b8ba07c3212715b763da56fd85", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "330d264a787121d5e2a6bf01", "sample_id": "61b5ac490087526a9f88f36e", "file": "20_seconds/HOT3D/61b5ac490087526a9f88f36e_MANO_GT_20.00s.mp4", "recording_id": "P0009_8c95667c", "source_first_frame": "1575", "source_last_frame": "2174", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 24/600, right absent 7/600; absence alone does not establish visible-hand label failure", "id": "330d264a787121d5e2a6bf01", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/61b5ac490087526a9f88f36e_MANO_GT_20.00s.mp4", "bytes": 6770981, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.14955767311654006, "p90": 0.3711932647801813, "valid_pairs": 599, "valid_duration_s": 19.966713752997748, "coverage": 1.0 }, "camera_rotation": { "median": 19.304643565341404, "p90": 47.62452288237668, "valid_pairs": 599, "valid_duration_s": 19.966713752997748, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.09655971217197648, "p90": 0.3604702574856747, "valid_pairs": 574, "valid_duration_s": 19.133381435996853, "coverage": 0.9582639222803612 }, "cmh_m_s": { "median": 0.14203881080888803, "p90": 0.4121715252013367, "valid_pairs": 574, "valid_duration_s": 19.133381435996853, "coverage": 0.9582639222803612 }, "rhcm_continuous_estimate_m_s": { "median": 0.139554055933543, "p90": 0.3511436693254836, "valid_pairs": 574, "valid_duration_s": 19.133381435996853, "coverage": 0.9582639222803612 }, "rhcm_discrete_m_s": { "median": 0.13911382349827345, "p90": 0.3517233555037389, "valid_pairs": 574, "valid_duration_s": 19.133381435996853, "coverage": 0.9582639222803612 }, "wms": { "median": 0.5902093274474962, "p90": 0.9434740934350178, "valid_pairs": 572, "valid_duration_s": 19.066723103002005, "coverage": 0.9549254493689218 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5144022443053605, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.17492785013390094, "p90": 0.7354593214252284, "valid_pairs": 592, "valid_duration_s": 19.73335094999493, "coverage": 0.9883124080462273 }, "cmh_m_s": { "median": 0.1538471108722991, "p90": 0.46999801925688967, "valid_pairs": 592, "valid_duration_s": 19.73335094999493, "coverage": 0.9883124080462273 }, "rhcm_continuous_estimate_m_s": { "median": 0.18325483580722512, "p90": 0.5841014144361689, "valid_pairs": 592, "valid_duration_s": 19.73335094999493, "coverage": 0.9883124080462273 }, "rhcm_discrete_m_s": { "median": 0.181908283674026, "p90": 0.5846024538118456, "valid_pairs": 592, "valid_duration_s": 19.73335094999493, "coverage": 0.9883124080462273 }, "wms": { "median": 0.5331981901922556, "p90": 0.9217306193241408, "valid_pairs": 591, "valid_duration_s": 19.700023435994808, "coverage": 0.9866432543531156 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4574296735337986, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.3711932647801813, "camera_angular_speed_p90_deg_s": 47.62452288237668, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.7354593214252284, "hand_camera_speed_p90_m_s": 0.5846024538118456, "hand_world_valid_fraction": 0.9741666666666666, "hand_camera_valid_fraction": 0.9741666666666666, "cmh_p90_m_s": 0.46999801925688967, "rhcm_p90_m_s": 0.5846024538118456, "rham_p90_deg_s": null, "wms_p90": 0.9434740934350178, "gt_coverage": 0.9883124080462273, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "53fd2430902eb9f853d1287529cf39b3608bd05a331c727cba2b2748fe57330f", "metadata_sha256": "1189f70b79b0b6d26f3390c5fa736551d2faf4313080e660be8ea2e504149dbc", "source_file": "long20_20260918/local/benchmark/previews/330d264a787121d5e2a6bf01/dense_reference.npz", "window_s": [ 41558.508181699, 41578.474895452 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b1a84ed9b29a3e91a60186e69807d7b9f0844c53cf51a907a47298adaafb5c28", "mano_sha256": "1a13e4d2bdf6ab087796c9be1a42483f4f078b155901f5025f92b9938da59b55", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "3497344f37452902bebb85cd", "sample_id": "2aa3234e3e6203a222b168a3", "file": "20_seconds/HOT3D/2aa3234e3e6203a222b168a3_MANO_GT_20.00s.mp4", "recording_id": "P0011_0c2d00ed", "source_first_frame": "141", "source_last_frame": "740", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "valid mesh not rasterized in 7 frames; per-hand official MANO coverage: left absent 42/600, right absent 0/600; absence alone does not establish visible-hand label failure", "id": "3497344f37452902bebb85cd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/2aa3234e3e6203a222b168a3_MANO_GT_20.00s.mp4", "bytes": 5172129, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.036304856235976415, "p90": 0.26893544332908875, "valid_pairs": 599, "valid_duration_s": 19.966689982997195, "coverage": 1.0 }, "camera_rotation": { "median": 13.869194288239061, "p90": 39.55669709358739, "valid_pairs": 599, "valid_duration_s": 19.966689982997195, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.019354230458812516, "p90": 0.29298681055915565, "valid_pairs": 544, "valid_duration_s": 18.133322399000463, "coverage": 0.9081786923341849 }, "cmh_m_s": { "median": 0.0867719280337596, "p90": 0.2736281767939479, "valid_pairs": 544, "valid_duration_s": 18.133322399000463, "coverage": 0.9081786923341849 }, "rhcm_continuous_estimate_m_s": { "median": 0.09517880266035782, "p90": 0.3477227988087773, "valid_pairs": 544, "valid_duration_s": 18.133322399000463, "coverage": 0.9081786923341849 }, "rhcm_discrete_m_s": { "median": 0.09509181060955625, "p90": 0.3479617304025115, "valid_pairs": 544, "valid_duration_s": 18.133322399000463, "coverage": 0.9081786923341849 }, "wms": { "median": 0.8350326203630122, "p90": 0.9712081380606818, "valid_pairs": 501, "valid_duration_s": 16.69992734600237, "coverage": 0.8363893745144204 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6712314359572202, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.0887443705021146, "p90": 0.5539785470512328, "valid_pairs": 599, "valid_duration_s": 19.966689982997195, "coverage": 1.0 }, "cmh_m_s": { "median": 0.09647079279221202, "p90": 0.4469858929056525, "valid_pairs": 599, "valid_duration_s": 19.966689982997195, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.10973751841908069, "p90": 0.42158962117101584, "valid_pairs": 599, "valid_duration_s": 19.966689982997195, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.10977841525225374, "p90": 0.42093110255319904, "valid_pairs": 599, "valid_duration_s": 19.966689982997195, "coverage": 1.0 }, "wms": { "median": 0.6081536442981548, "p90": 0.8669837242942297, "valid_pairs": 597, "valid_duration_s": 19.90000158100156, "coverage": 0.9966600171559521 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6629132723993724, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.26893544332908875, "camera_angular_speed_p90_deg_s": 39.55669709358739, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5539785470512328, "hand_camera_speed_p90_m_s": 0.42093110255319904, "hand_world_valid_fraction": 0.9558333333333333, "hand_camera_valid_fraction": 0.9558333333333333, "cmh_p90_m_s": 0.4469858929056525, "rhcm_p90_m_s": 0.42093110255319904, "rham_p90_deg_s": null, "wms_p90": 0.9712081380606818, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "ed376ec1aff49bcb3565831dec083e057453cf188b41966af11667e1cc0e7902", "metadata_sha256": "0b2e705145206863622534ac79f308465b7cafb44004779c4d2a1772021ccca7", "source_file": "long20_20260918/local/train/previews/3497344f37452902bebb85cd/dense_reference.npz", "window_s": [ 55564.708196558, 55584.674886541 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "92f10e7cd9ccfa8032278ec6a919873349b35c25735a6cf62c03190ac9da40c3", "mano_sha256": "8f6bd06deddcf822d355fbf04e9bf105a2dc0ee3d803d7d44a982e9e3669da6e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "3e37280131ee83e74a1a11be", "sample_id": "626286517de6f4b610e7b06a", "file": "20_seconds/HOT3D/626286517de6f4b610e7b06a_MANO_GT_20.00s.mp4", "recording_id": "P0009_5bb7cc2a", "source_first_frame": "525", "source_last_frame": "1124", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "", "id": "3e37280131ee83e74a1a11be", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/626286517de6f4b610e7b06a_MANO_GT_20.00s.mp4", "bytes": 6800410, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.09219904670637373, "p90": 0.420825522905464, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "camera_rotation": { "median": 15.60529123329689, "p90": 35.802943245236285, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.11915235730757158, "p90": 0.5111147148472723, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "cmh_m_s": { "median": 0.11755339890285611, "p90": 0.4888684719443828, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.15144706807513766, "p90": 0.47812964629428806, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.15147611396290983, "p90": 0.47752211196611005, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "wms": { "median": 0.6064374187428119, "p90": 0.9269923451507103, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5762631699432162, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.15599001927758785, "p90": 0.7043819650561165, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "cmh_m_s": { "median": 0.12151949182702704, "p90": 0.4234162767193611, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.1707666734109821, "p90": 0.6051941990233762, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.17093682606648836, "p90": 0.6047357839701528, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "wms": { "median": 0.6529527450723299, "p90": 0.9139917007797157, "valid_pairs": 599, "valid_duration_s": 19.966598179002176, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4618356123081389, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.420825522905464, "camera_angular_speed_p90_deg_s": 35.802943245236285, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.7043819650561165, "hand_camera_speed_p90_m_s": 0.6047357839701528, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.4888684719443828, "rhcm_p90_m_s": 0.6047357839701528, "rham_p90_deg_s": null, "wms_p90": 0.9269923451507103, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "316068694d6745f80ec7c5fcd8f19603d3a4fb2b8431e37e00c52cae00c54cbc", "metadata_sha256": "aca00eb6ffc321f01e331b11485c18657b94162ccc3239dc84db8d256a32f765", "source_file": "long20_20260918/local/benchmark/previews/3e37280131ee83e74a1a11be/dense_reference.npz", "window_s": [ 41236.508278449, 41256.474876628 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "174041be8e137b2acbce4b69fda150cb9ce2714eabff850fad23a18922807f67", "mano_sha256": "d756fc458e132b420b99e7cd0cd5f734010d2d2897061afae8fb8ac8f405281d", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "3f87381250908cfb48c6beda", "sample_id": "1c28111d6422e32551eac580", "file": "20_seconds/HOT3D/1c28111d6422e32551eac580_MANO_GT_20.00s.mp4", "recording_id": "P0009_0e02bf39", "source_first_frame": "2625", "source_last_frame": "3224", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "", "id": "3f87381250908cfb48c6beda", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/1c28111d6422e32551eac580_MANO_GT_20.00s.mp4", "bytes": 7027923, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.13032344887918598, "p90": 0.4399017290603013, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "camera_rotation": { "median": 18.839138110962647, "p90": 48.65432929982302, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.13824774699852663, "p90": 0.5139680826358527, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "cmh_m_s": { "median": 0.14550945986614336, "p90": 0.38764991691044337, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.11815326204179355, "p90": 0.49155500313988937, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.11822332879059207, "p90": 0.49420995758220954, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "wms": { "median": 0.4140932545563007, "p90": 0.8940769082385924, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6820449913735823, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.12587417293536957, "p90": 0.6105158036060894, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "cmh_m_s": { "median": 0.15208214177639814, "p90": 0.40479893934468275, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.14226851177375438, "p90": 0.5176712091859644, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.1423195226414118, "p90": 0.5163517275476068, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "wms": { "median": 0.5093558584019425, "p90": 0.8977814178181741, "valid_pairs": 599, "valid_duration_s": 19.966683712002123, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6249924902708527, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.4399017290603013, "camera_angular_speed_p90_deg_s": 48.65432929982302, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6105158036060894, "hand_camera_speed_p90_m_s": 0.5163517275476068, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.40479893934468275, "rhcm_p90_m_s": 0.5163517275476068, "rham_p90_deg_s": null, "wms_p90": 0.8977814178181741, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "d0d9f873a9500180fedcaca820cb51ee2daf696778db8107826c6307532db09a", "metadata_sha256": "d5c1476272aa5372b7755db12ac249b6d92616d5c8f20435909142dd0137624a", "source_file": "long20_20260918/local/benchmark/previews/3f87381250908cfb48c6beda/dense_reference.npz", "window_s": [ 39762.508189059, 39782.474872771 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "bc6a5fbdc849c643f5619830bdcddef6804894cf46f46274e4e738fd20bd7815", "mano_sha256": "5eaf3abc20348cef49df82a34543b103ea45515479cf1e0fb6e753319b112e52", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "439bffb32b5e573a4a75e4b8", "sample_id": "efaf034af387a2b20f36182f", "file": "20_seconds/HOT3D/efaf034af387a2b20f36182f_MANO_GT_20.00s.mp4", "recording_id": "P0011_47878e48", "source_first_frame": "2315", "source_last_frame": "2914", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 18/600, right absent 0/600; absence alone does not establish visible-hand label failure", "id": "439bffb32b5e573a4a75e4b8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/efaf034af387a2b20f36182f_MANO_GT_20.00s.mp4", "bytes": 5493847, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.056252114367709355, "p90": 0.14092823535539498, "valid_pairs": 599, "valid_duration_s": 19.96668350500113, "coverage": 1.0 }, "camera_rotation": { "median": 14.716030692172101, "p90": 35.16582617846735, "valid_pairs": 599, "valid_duration_s": 19.96668350500113, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06640450137962561, "p90": 0.22826060524778316, "valid_pairs": 580, "valid_duration_s": 19.333354421003605, "coverage": 0.9682807070168218 }, "cmh_m_s": { "median": 0.0982102814105309, "p90": 0.28642838726603786, "valid_pairs": 580, "valid_duration_s": 19.333354421003605, "coverage": 0.9682807070168218 }, "rhcm_continuous_estimate_m_s": { "median": 0.1134368904449645, "p90": 0.362044314151904, "valid_pairs": 580, "valid_duration_s": 19.333354421003605, "coverage": 0.9682807070168218 }, "rhcm_discrete_m_s": { "median": 0.11384714319356015, "p90": 0.3622297781049641, "valid_pairs": 580, "valid_duration_s": 19.333354421003605, "coverage": 0.9682807070168218 }, "wms": { "median": 0.7518842754615056, "p90": 0.9491688066949376, "valid_pairs": 575, "valid_duration_s": 19.16667608400894, "coverage": 0.9599328841571607 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4829796636198947, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08294371771551806, "p90": 0.47776263020326887, "valid_pairs": 599, "valid_duration_s": 19.96668350500113, "coverage": 1.0 }, "cmh_m_s": { "median": 0.1004778345338136, "p90": 0.2812090834408813, "valid_pairs": 599, "valid_duration_s": 19.96668350500113, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.1297526254873768, "p90": 0.4172754284264862, "valid_pairs": 599, "valid_duration_s": 19.96668350500113, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.12960213116154387, "p90": 0.41605186399159344, "valid_pairs": 599, "valid_duration_s": 19.96668350500113, "coverage": 1.0 }, "wms": { "median": 0.7144989911838524, "p90": 0.9347185193925447, "valid_pairs": 592, "valid_duration_s": 19.733357810997404, "coverage": 0.9883142488863871 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4917462772402163, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.14092823535539498, "camera_angular_speed_p90_deg_s": 35.16582617846735, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.47776263020326887, "hand_camera_speed_p90_m_s": 0.41605186399159344, "hand_world_valid_fraction": 0.985, "hand_camera_valid_fraction": 0.985, "cmh_p90_m_s": 0.28642838726603786, "rhcm_p90_m_s": 0.41605186399159344, "rham_p90_deg_s": null, "wms_p90": 0.9491688066949376, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "a42f32c3d8e5c6f5a93147c4fcf2a5d3075e45a6fee969ef166cdee5beb76e7a", "metadata_sha256": "dc469cbf1cb8ccff83373e7bc28b93bc627ec726d9eb32bc6dda165f133d4dc3", "source_file": "long20_20260918/local/train/previews/439bffb32b5e573a4a75e4b8/dense_reference.npz", "window_s": [ 57659.341523311, 57679.308206816 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7ef6f872b4c277891bf7266ea7f09cae3630a2c542622624e6e2050a1bf70993", "mano_sha256": "ad9668eeb497a35247ece3d568b3b55dd08d4d78e67629d66de8f2fa16bcc7bf", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "4e2b6365dd1c2a7bf3328ed9", "sample_id": "83907e832c8794973d8df9d6", "file": "20_seconds/HOT3D/83907e832c8794973d8df9d6_MANO_GT_20.00s.mp4", "recording_id": "P0009_1b21bb01", "source_first_frame": "3156", "source_last_frame": "3755", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 11/600 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 14 frames; per-hand official MANO coverage: left absent 0/600, right absent 73/600; absence alone does not establish visible-hand label failure", "id": "4e2b6365dd1c2a7bf3328ed9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/83907e832c8794973d8df9d6_MANO_GT_20.00s.mp4", "bytes": 5437334, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08228142201237867, "p90": 0.3613969561151813, "valid_pairs": 599, "valid_duration_s": 19.9666573459981, "coverage": 1.0 }, "camera_rotation": { "median": 15.958441998004405, "p90": 62.52656638210451, "valid_pairs": 599, "valid_duration_s": 19.9666573459981, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.22353590122030156, "p90": 0.51903002221883, "valid_pairs": 571, "valid_duration_s": 19.033319679998385, "coverage": 0.9532551868935246 }, "cmh_m_s": { "median": 0.13245953219148587, "p90": 0.6039242567502048, "valid_pairs": 571, "valid_duration_s": 19.033319679998385, "coverage": 0.9532551868935246 }, "rhcm_continuous_estimate_m_s": { "median": 0.2755829472800691, "p90": 0.572172751571813, "valid_pairs": 571, "valid_duration_s": 19.033319679998385, "coverage": 0.9532551868935246 }, "rhcm_discrete_m_s": { "median": 0.2766970470857929, "p90": 0.5704752706083337, "valid_pairs": 571, "valid_duration_s": 19.033319679998385, "coverage": 0.9532551868935246 }, "wms": { "median": 0.7138002601978098, "p90": 0.9524472901207047, "valid_pairs": 571, "valid_duration_s": 19.033319679998385, "coverage": 0.9532551868935246 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.14433620577965678, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06804553091590661, "p90": 0.5043477184366447, "valid_pairs": 525, "valid_duration_s": 17.499987356997735, "coverage": 0.876460543882937 }, "cmh_m_s": { "median": 0.10517239534787845, "p90": 0.47194124623722244, "valid_pairs": 525, "valid_duration_s": 17.499987356997735, "coverage": 0.876460543882937 }, "rhcm_continuous_estimate_m_s": { "median": 0.13893115753804944, "p90": 0.5493738810026355, "valid_pairs": 525, "valid_duration_s": 17.499987356997735, "coverage": 0.876460543882937 }, "rhcm_discrete_m_s": { "median": 0.13972594850744496, "p90": 0.5422586333544295, "valid_pairs": 525, "valid_duration_s": 17.499987356997735, "coverage": 0.876460543882937 }, "wms": { "median": 0.7890418477708419, "p90": 0.9707187971655561, "valid_pairs": 524, "valid_duration_s": 17.466659774996515, "coverage": 0.8747913820686336 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.32532236462924646, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.3613969561151813, "camera_angular_speed_p90_deg_s": 62.52656638210451, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.51903002221883, "hand_camera_speed_p90_m_s": 0.5704752706083337, "hand_world_valid_fraction": 0.9166666666666666, "hand_camera_valid_fraction": 0.9166666666666666, "cmh_p90_m_s": 0.6039242567502048, "rhcm_p90_m_s": 0.5704752706083337, "rham_p90_deg_s": null, "wms_p90": 0.9707187971655561, "gt_coverage": 0.9532551868935246, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "b758cecb9a286af729fbdf4b32d91354f2af014e556ea9ec65788febcf86fa17", "metadata_sha256": "6c22cc682dd3dad09a52525155d58944baf6a36cc1ad5c136b23f5956b4fa502", "source_file": "long20_20260918/local/benchmark/previews/4e2b6365dd1c2a7bf3328ed9/dense_reference.npz", "window_s": [ 44424.208200499, 44444.174857845 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "906ea62e4e5c92aa1fe1f5678884855d1ba3e905ce593be44a02b66b11324bc9", "mano_sha256": "d16f05dc9c6b471b9aa15ce8ca49447efef0809a1d78d5514465d7b23f01f695", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "53882cf41f01496533f90326", "sample_id": "1d00261e162ef3bb59278794", "file": "20_seconds/HOT3D/1d00261e162ef3bb59278794_MANO_GT_20.00s.mp4", "recording_id": "P0009_927973d7", "source_first_frame": "3075", "source_last_frame": "3674", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 9/600 output frames; may include genuine out-of-view frames; per-hand official MANO coverage: left absent 45/600, right absent 9/600; absence alone does not establish visible-hand label failure", "id": "53882cf41f01496533f90326", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/1d00261e162ef3bb59278794_MANO_GT_20.00s.mp4", "bytes": 7094458, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08391808909520349, "p90": 0.3870158916393594, "valid_pairs": 599, "valid_duration_s": 19.96665068299626, "coverage": 1.0 }, "camera_rotation": { "median": 14.272462581151535, "p90": 41.34608900689058, "valid_pairs": 599, "valid_duration_s": 19.96665068299626, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.13323592366033735, "p90": 0.45816003646634523, "valid_pairs": 553, "valid_duration_s": 18.433300046999648, "coverage": 0.9232044141833751 }, "cmh_m_s": { "median": 0.10394277608101185, "p90": 0.3517743002195211, "valid_pairs": 553, "valid_duration_s": 18.433300046999648, "coverage": 0.9232044141833751 }, "rhcm_continuous_estimate_m_s": { "median": 0.13790511994845028, "p90": 0.41197909558000423, "valid_pairs": 553, "valid_duration_s": 18.433300046999648, "coverage": 0.9232044141833751 }, "rhcm_discrete_m_s": { "median": 0.1378222224309, "p90": 0.41121264215077247, "valid_pairs": 553, "valid_duration_s": 18.433300046999648, "coverage": 0.9232044141833751 }, "wms": { "median": 0.5581003641484061, "p90": 0.8325074855811158, "valid_pairs": 552, "valid_duration_s": 18.39996511299978, "coverage": 0.921534883598145 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.7229876250403402, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1227411785468194, "p90": 0.5247141056429012, "valid_pairs": 589, "valid_duration_s": 19.633304450995638, "coverage": 0.9833048498071586 }, "cmh_m_s": { "median": 0.10963146923827334, "p90": 0.38578943822066225, "valid_pairs": 589, "valid_duration_s": 19.633304450995638, "coverage": 0.9833048498071586 }, "rhcm_continuous_estimate_m_s": { "median": 0.12622639937418204, "p90": 0.4777702517982198, "valid_pairs": 589, "valid_duration_s": 19.633304450995638, "coverage": 0.9833048498071586 }, "rhcm_discrete_m_s": { "median": 0.12622624809491456, "p90": 0.4743613920945988, "valid_pairs": 589, "valid_duration_s": 19.633304450995638, "coverage": 0.9833048498071586 }, "wms": { "median": 0.6082658329624424, "p90": 0.8703322155898514, "valid_pairs": 589, "valid_duration_s": 19.633304450995638, "coverage": 0.9833048498071586 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6234165042380221, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.3870158916393594, "camera_angular_speed_p90_deg_s": 41.34608900689058, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5247141056429012, "hand_camera_speed_p90_m_s": 0.4743613920945988, "hand_world_valid_fraction": 0.955, "hand_camera_valid_fraction": 0.955, "cmh_p90_m_s": 0.38578943822066225, "rhcm_p90_m_s": 0.4743613920945988, "rham_p90_deg_s": null, "wms_p90": 0.8703322155898514, "gt_coverage": 0.9833048498071586, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "80a435500c2de11814f8837b7359cbe1b52303dc99edda8aa60faa5fcc2727c8", "metadata_sha256": "c0e9edae6e8a4f3ec248ad95badb8c584f94ee342ed954ee4e4e602e8a8abfbe", "source_file": "long20_20260918/local/benchmark/previews/53882cf41f01496533f90326/dense_reference.npz", "window_s": [ 40107.508190038, 40127.474840721 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "1a5d6328af38bc5b657f71f5d47826d3618e6f1477b27550447c7c871f1f6e5b", "mano_sha256": "e2c870cc368d79fbe86133da95d45a491854fb843d106cb4165f1f6b1b60f44a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "74f15a7f3b66289a17f78eba", "sample_id": "fcbc130d3c3340fee3ec4906", "file": "20_seconds/HOT3D/fcbc130d3c3340fee3ec4906_MANO_GT_20.00s.mp4", "recording_id": "P0009_9e03121a", "source_first_frame": "525", "source_last_frame": "1124", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 0/600, right absent 20/600; absence alone does not establish visible-hand label failure", "id": "74f15a7f3b66289a17f78eba", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/fcbc130d3c3340fee3ec4906_MANO_GT_20.00s.mp4", "bytes": 6558513, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.07300236088880041, "p90": 0.30870400941972514, "valid_pairs": 599, "valid_duration_s": 19.966661568003474, "coverage": 1.0 }, "camera_rotation": { "median": 14.37515445403235, "p90": 50.821711021293474, "valid_pairs": 599, "valid_duration_s": 19.966661568003474, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.17075836907449266, "p90": 0.5585411199678159, "valid_pairs": 586, "valid_duration_s": 19.533321111004625, "coverage": 0.978296799616553 }, "cmh_m_s": { "median": 0.10497260688936641, "p90": 0.48559745585125935, "valid_pairs": 586, "valid_duration_s": 19.533321111004625, "coverage": 0.978296799616553 }, "rhcm_continuous_estimate_m_s": { "median": 0.22547412483595092, "p90": 0.6815340201233371, "valid_pairs": 586, "valid_duration_s": 19.533321111004625, "coverage": 0.978296799616553 }, "rhcm_discrete_m_s": { "median": 0.22542298203369948, "p90": 0.6798166261769609, "valid_pairs": 586, "valid_duration_s": 19.533321111004625, "coverage": 0.978296799616553 }, "wms": { "median": 0.7346420817094532, "p90": 0.9294477606760695, "valid_pairs": 586, "valid_duration_s": 19.533321111004625, "coverage": 0.978296799616553 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.17214243023151626, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.12183365226509574, "p90": 0.5283623852368459, "valid_pairs": 578, "valid_duration_s": 19.266627864999464, "coverage": 0.9649398723657533 }, "cmh_m_s": { "median": 0.10633950336371498, "p90": 0.4853295782414574, "valid_pairs": 578, "valid_duration_s": 19.266627864999464, "coverage": 0.9649398723657533 }, "rhcm_continuous_estimate_m_s": { "median": 0.16985623063554933, "p90": 0.6103498249610176, "valid_pairs": 578, "valid_duration_s": 19.266627864999464, "coverage": 0.9649398723657533 }, "rhcm_discrete_m_s": { "median": 0.17009634817765423, "p90": 0.6094006085062387, "valid_pairs": 578, "valid_duration_s": 19.266627864999464, "coverage": 0.9649398723657533 }, "wms": { "median": 0.7085659656385734, "p90": 0.9405673860617875, "valid_pairs": 576, "valid_duration_s": 19.19996926199383, "coverage": 0.9616013772057784 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.37323357692425163, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.30870400941972514, "camera_angular_speed_p90_deg_s": 50.821711021293474, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5585411199678159, "hand_camera_speed_p90_m_s": 0.6798166261769609, "hand_world_valid_fraction": 0.9733333333333334, "hand_camera_valid_fraction": 0.9733333333333334, "cmh_p90_m_s": 0.48559745585125935, "rhcm_p90_m_s": 0.6798166261769609, "rham_p90_deg_s": null, "wms_p90": 0.9405673860617875, "gt_coverage": 0.978296799616553, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "e9cff50994e73f76427bc83eb9e741460646ac57760d0e0d04c0d0e63e3203c6", "metadata_sha256": "934175cc597a9ca72c2b2b389cbd4ca389eaf2e5573353a255d88326a996112b", "source_file": "long20_20260918/local/benchmark/previews/74f15a7f3b66289a17f78eba/dense_reference.npz", "window_s": [ 43686.508197616, 43706.474859184 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5347496db00eabb91800575f0887caec3c2c48ef934b8f555b6fdca6f53fe927", "mano_sha256": "3c0d008809869b689f35a79bccd3aeddc7c8e1493f1fe8c2d3458bb611df2c7b", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "76f172a0f08564f96a1dfe6e", "sample_id": "50e6822cb19ce8605d7abb92", "file": "20_seconds/HOT3D/50e6822cb19ce8605d7abb92_MANO_GT_20.00s.mp4", "recording_id": "P0012_ca1f6626", "source_first_frame": "2299", "source_last_frame": "2898", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "76f172a0f08564f96a1dfe6e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/50e6822cb19ce8605d7abb92_MANO_GT_20.00s.mp4", "bytes": 7035607, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.09890784288873024, "p90": 0.264685908822553, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "camera_rotation": { "median": 17.54209299427609, "p90": 43.24903714255478, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.15641655086443326, "p90": 0.6309770989183924, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "cmh_m_s": { "median": 0.12598794058937568, "p90": 0.410397035699982, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.14284706230913377, "p90": 0.5577310530498165, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.14289357533903246, "p90": 0.5570537540958578, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "wms": { "median": 0.5640239487476794, "p90": 0.8437375008016591, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.674297072601493, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.14203226883011477, "p90": 0.5387538129638956, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "cmh_m_s": { "median": 0.12135486797957443, "p90": 0.2882376483442858, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.09637630644359454, "p90": 0.4942981510807216, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0968883112440674, "p90": 0.4939374636056675, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "wms": { "median": 0.4698767716199887, "p90": 0.8713481061742278, "valid_pairs": 599, "valid_duration_s": 19.966732159002277, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.7503326056247591, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.264685908822553, "camera_angular_speed_p90_deg_s": 43.24903714255478, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6309770989183924, "hand_camera_speed_p90_m_s": 0.5570537540958578, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.410397035699982, "rhcm_p90_m_s": 0.5570537540958578, "rham_p90_deg_s": null, "wms_p90": 0.8713481061742278, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "8e4529b12823d9860d6a13c84031f2ebbed64898a3865cc19bd528ee8d8f276a", "metadata_sha256": "59b150ab5d8f87cd6ddfe2db124fe344506f4338008fb76c58f206b6421df352", "source_file": "long20_20260918/local/train/previews/76f172a0f08564f96a1dfe6e/dense_reference.npz", "window_s": [ 57336.64153209, 57356.608264249 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "34d9399aadfb0c5e933674a5af1303bf3a6c327fffccecc14d6b9bc5a4155ced", "mano_sha256": "86dacd2bb25c26effd47a8856b99852b0649dd017c55bc4b844b37ac499039b8", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "78e7dfc11f23ae32a2269158", "sample_id": "1df99420ba4493cac5248741", "file": "20_seconds/HOT3D/1df99420ba4493cac5248741_MANO_GT_20.00s.mp4", "recording_id": "P0011_ccc678c7", "source_first_frame": "1606", "source_last_frame": "2205", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "valid mesh not rasterized in 4 frames", "id": "78e7dfc11f23ae32a2269158", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/1df99420ba4493cac5248741_MANO_GT_20.00s.mp4", "bytes": 5846623, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.10176058893366237, "p90": 0.34040482029425806, "valid_pairs": 599, "valid_duration_s": 19.966642785999284, "coverage": 1.0 }, "camera_rotation": { "median": 13.990343751849963, "p90": 41.247430231910954, "valid_pairs": 599, "valid_duration_s": 19.966642785999284, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.009650433740265434, "p90": 0.04025111605057357, "valid_pairs": 553, "valid_duration_s": 18.433332273001724, "coverage": 0.9232063933115123 }, "cmh_m_s": { "median": 0.09914519897811978, "p90": 0.2944851273994727, "valid_pairs": 553, "valid_duration_s": 18.433332273001724, "coverage": 0.9232063933115123 }, "rhcm_continuous_estimate_m_s": { "median": 0.0921384674000619, "p90": 0.29150411698162215, "valid_pairs": 553, "valid_duration_s": 18.433332273001724, "coverage": 0.9232063933115123 }, "rhcm_discrete_m_s": { "median": 0.09264096949925775, "p90": 0.2940529422555234, "valid_pairs": 553, "valid_duration_s": 18.433332273001724, "coverage": 0.9232063933115123 }, "wms": { "median": 0.870089348732036, "p90": 0.9722485149897716, "valid_pairs": 512, "valid_duration_s": 17.06662135499937, "coverage": 0.8547566828293526 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5961675630676186, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.14155291299399322, "p90": 0.6051263326827143, "valid_pairs": 599, "valid_duration_s": 19.966642785999284, "coverage": 1.0 }, "cmh_m_s": { "median": 0.13551279444815445, "p90": 0.4931521264053341, "valid_pairs": 599, "valid_duration_s": 19.966642785999284, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.12187364843546687, "p90": 0.4902035469050833, "valid_pairs": 599, "valid_duration_s": 19.966642785999284, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.1215647277961174, "p90": 0.49101384297626277, "valid_pairs": 599, "valid_duration_s": 19.966642785999284, "coverage": 1.0 }, "wms": { "median": 0.5427069430418584, "p90": 0.8847806824049478, "valid_pairs": 599, "valid_duration_s": 19.966642785999284, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6631114845813255, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.34040482029425806, "camera_angular_speed_p90_deg_s": 41.247430231910954, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6051263326827143, "hand_camera_speed_p90_m_s": 0.49101384297626277, "hand_world_valid_fraction": 0.9625, "hand_camera_valid_fraction": 0.9625, "cmh_p90_m_s": 0.4931521264053341, "rhcm_p90_m_s": 0.49101384297626277, "rham_p90_deg_s": null, "wms_p90": 0.9722485149897716, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "4f101d94ce2128f2966250b8d855121bd7f7c52cbc278468b3d9b5ecf0c4ff9f", "metadata_sha256": "59443cdfcb6e5ece2681030b07a33d78f0d1c275fa7f947caf6fbf115dfba84a", "source_file": "long20_20260918/local/train/previews/78e7dfc11f23ae32a2269158/dense_reference.npz", "window_s": [ 55922.541546132, 55942.508188918 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b0a3243c5b10a0221876741faaa5940bb362230990298a8d35a77eddb5579c25", "mano_sha256": "5d3f65f256095d13132b4bb9c3dc7bedd52b30efca6692a8f93e74a7f4c394a2", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "7a3b5530706123c0bf7709cc", "sample_id": "2963491c7ce140f87c3e88d3", "file": "20_seconds/HOT3D/2963491c7ce140f87c3e88d3_MANO_GT_20.00s.mp4", "recording_id": "P0014_0de48e2c", "source_first_frame": "267", "source_last_frame": "866", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "7a3b5530706123c0bf7709cc", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/2963491c7ce140f87c3e88d3_MANO_GT_20.00s.mp4", "bytes": 5709386, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02840178088037815, "p90": 0.13720379761625023, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "camera_rotation": { "median": 7.669044827312271, "p90": 20.690307397426587, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06698722423531729, "p90": 0.15961481397067132, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "cmh_m_s": { "median": 0.05409664673145298, "p90": 0.15874131739697633, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.07923738938562833, "p90": 0.20867350919431646, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.07920196385647361, "p90": 0.20860028584449147, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "wms": { "median": 0.6688240061254058, "p90": 0.8997157080040736, "valid_pairs": 583, "valid_duration_s": 19.433248621993698, "coverage": 0.9732861878545276 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4912132818912233, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.11316584607038221, "p90": 0.39088817992769737, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "cmh_m_s": { "median": 0.060410192504171525, "p90": 0.17583769198937604, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.12039720310072367, "p90": 0.3423325767799648, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.12005738035663328, "p90": 0.3412704078023016, "valid_pairs": 599, "valid_duration_s": 19.966633518997696, "coverage": 1.0 }, "wms": { "median": 0.6729488240521133, "p90": 0.9426427465960242, "valid_pairs": 592, "valid_duration_s": 19.7333109780011, "coverage": 0.9883143775451882 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5256849487160875, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.13720379761625023, "camera_angular_speed_p90_deg_s": 20.690307397426587, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.39088817992769737, "hand_camera_speed_p90_m_s": 0.3412704078023016, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.17583769198937604, "rhcm_p90_m_s": 0.3412704078023016, "rham_p90_deg_s": null, "wms_p90": 0.9426427465960242, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "272f733e4b8a06989ced0e547d9976e9fcb8fc111b45a4e8030de1946e9cb962", "metadata_sha256": "8c31ab3a5537047f086a65cc6fe44c86b88f1105c7572376d8811cfbc8d0a664", "source_file": "long20_20260918/local/train/previews/7a3b5530706123c0bf7709cc/dense_reference.npz", "window_s": [ 42191.908228161, 42211.87486168 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "272dbf07031a31392da1f6b3b22a076b8e94f9812d70e30c4f80a1fe25fcf0d5", "mano_sha256": "416e664a82095e4a15e5a20f1285357410a1316dd7bfec72c269c55785a3783d", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "7d8e7e75cbb40eac2b16f63c", "sample_id": "8877c88c900601263996d887", "file": "20_seconds/HOT3D/8877c88c900601263996d887_MANO_GT_20.00s.mp4", "recording_id": "P0012_44c5f677", "source_first_frame": "2963", "source_last_frame": "3562", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "7d8e7e75cbb40eac2b16f63c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/8877c88c900601263996d887_MANO_GT_20.00s.mp4", "bytes": 7759232, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.18779501855842734, "p90": 0.3464691168000391, "valid_pairs": 598, "valid_duration_s": 19.933347536003566, "coverage": 0.99666725455885 }, "camera_rotation": { "median": 24.7304487915968, "p90": 94.35943499015701, "valid_pairs": 598, "valid_duration_s": 19.933347536003566, "coverage": 0.99666725455885 }, "sides": { "left": { "whm_m_s": { "median": 0.2665723237027732, "p90": 0.7524092445866813, "valid_pairs": 598, "valid_duration_s": 19.933347536003566, "coverage": 0.99666725455885 }, "cmh_m_s": { "median": 0.22413539240585711, "p90": 0.6956844505176657, "valid_pairs": 598, "valid_duration_s": 19.933347536003566, "coverage": 0.99666725455885 }, "rhcm_continuous_estimate_m_s": { "median": 0.23240618212004469, "p90": 0.8628140676028844, "valid_pairs": 598, "valid_duration_s": 19.933347536003566, "coverage": 0.99666725455885 }, "rhcm_discrete_m_s": { "median": 0.23138812806589154, "p90": 0.8636773909859269, "valid_pairs": 598, "valid_duration_s": 19.933347536003566, "coverage": 0.99666725455885 }, "wms": { "median": 0.5075387915805714, "p90": 0.8783233493039899, "valid_pairs": 598, "valid_duration_s": 19.933347536003566, "coverage": 0.99666725455885 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6062405931983185, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.29018482085033526, "p90": 0.7235506945356841, "valid_pairs": 413, "valid_duration_s": 13.766579274008109, "coverage": 0.6883288792768065 }, "cmh_m_s": { "median": 0.2228469308606701, "p90": 0.6601693315569319, "valid_pairs": 413, "valid_duration_s": 13.766579274008109, "coverage": 0.6883288792768065 }, "rhcm_continuous_estimate_m_s": { "median": 0.3763809992001251, "p90": 0.8807837686107016, "valid_pairs": 413, "valid_duration_s": 13.766579274008109, "coverage": 0.6883288792768065 }, "rhcm_discrete_m_s": { "median": 0.37586484290391714, "p90": 0.8834772769923455, "valid_pairs": 413, "valid_duration_s": 13.766579274008109, "coverage": 0.6883288792768065 }, "wms": { "median": 0.6553764519054709, "p90": 0.9143152606549791, "valid_pairs": 413, "valid_duration_s": 13.766579274008109, "coverage": 0.6883288792768065 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.17798381774814656, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.3464691168000391, "camera_angular_speed_p90_deg_s": 94.35943499015701, "camera_valid_fraction": 1.0, "camera_pairs": 598, "hand_world_speed_p90_m_s": 0.7524092445866813, "hand_camera_speed_p90_m_s": 0.8834772769923455, "hand_world_valid_fraction": 0.8483333333333334, "hand_camera_valid_fraction": 0.8483333333333334, "cmh_p90_m_s": 0.6956844505176657, "rhcm_p90_m_s": 0.8834772769923455, "rham_p90_deg_s": null, "wms_p90": 0.9143152606549791, "gt_coverage": 0.99666725455885, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "49d64a63c5e239e5263e10df6eb62adb1f9075ff63aa9a564898b1e8b9d15d7f", "metadata_sha256": "f0b1c1a2b2c559f4bde613cfa74549aba1f8fbfef426e3b1aba1b6ba6713f682", "source_file": "long20_20260918/local/train/previews/7d8e7e75cbb40eac2b16f63c/dense_reference.npz", "window_s": [ 59419.77485112, 59439.774853573 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8a8a0d3481863b723cce467a87518a821d533e05417e3bdb16c136cd6055755c", "mano_sha256": "e2054c946b99f278015e53aef4f8be92496b5f7bd6bea821b64e76a8591accce", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "7e7c8d899a311901ef84edd0", "sample_id": "1a27b45a1aec069b7b8898ae", "file": "20_seconds/HOT3D/1a27b45a1aec069b7b8898ae_MANO_GT_20.00s.mp4", "recording_id": "P0009_24af5475", "source_first_frame": "975", "source_last_frame": "1574", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "", "id": "7e7c8d899a311901ef84edd0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/1a27b45a1aec069b7b8898ae_MANO_GT_20.00s.mp4", "bytes": 5133475, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08549802609435395, "p90": 0.22569236625753816, "valid_pairs": 599, "valid_duration_s": 19.966649750000215, "coverage": 1.0 }, "camera_rotation": { "median": 14.801167849676718, "p90": 58.25826055294037, "valid_pairs": 599, "valid_duration_s": 19.966649750000215, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.10888994972154634, "p90": 0.7791084422819303, "valid_pairs": 514, "valid_duration_s": 17.133400273000007, "coverage": 0.8581009076397417 }, "cmh_m_s": { "median": 0.14061678976497044, "p90": 0.5137183668907518, "valid_pairs": 514, "valid_duration_s": 17.133400273000007, "coverage": 0.8581009076397417 }, "rhcm_continuous_estimate_m_s": { "median": 0.21426217410450815, "p90": 0.8271216254860251, "valid_pairs": 514, "valid_duration_s": 17.133400273000007, "coverage": 0.8581009076397417 }, "rhcm_discrete_m_s": { "median": 0.21422288640890325, "p90": 0.8274795006966994, "valid_pairs": 514, "valid_duration_s": 17.133400273000007, "coverage": 0.8581009076397417 }, "wms": { "median": 0.7462535518949652, "p90": 0.941152620665202, "valid_pairs": 511, "valid_duration_s": 17.033370634992025, "coverage": 0.853091071775416 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.3284352886674395, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.2018265359304725, "p90": 0.7679543733490046, "valid_pairs": 599, "valid_duration_s": 19.966649750000215, "coverage": 1.0 }, "cmh_m_s": { "median": 0.12290650617842329, "p90": 0.3831376699856903, "valid_pairs": 599, "valid_duration_s": 19.966649750000215, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.24546474779642632, "p90": 0.8102378408878326, "valid_pairs": 599, "valid_duration_s": 19.966649750000215, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.2442310096224915, "p90": 0.8101958044148234, "valid_pairs": 599, "valid_duration_s": 19.966649750000215, "coverage": 1.0 }, "wms": { "median": 0.7294246243773732, "p90": 0.9265283755147112, "valid_pairs": 598, "valid_duration_s": 19.9333297819976, "coverage": 0.9983312188865027 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.408335129340428, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.22569236625753816, "camera_angular_speed_p90_deg_s": 58.25826055294037, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.7791084422819303, "hand_camera_speed_p90_m_s": 0.8274795006966994, "hand_world_valid_fraction": 0.93, "hand_camera_valid_fraction": 0.93, "cmh_p90_m_s": 0.5137183668907518, "rhcm_p90_m_s": 0.8274795006966994, "rham_p90_deg_s": null, "wms_p90": 0.941152620665202, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "4bcb906d92d07cf15b3e7328c40b49ab2e837ab72131e7727d35496cc3101d78", "metadata_sha256": "4e79db8994323fc5cc93e3f7adffd71ecf50aefda4b63104b79f8a840834c58d", "source_file": "long20_20260918/local/benchmark/previews/7e7c8d899a311901ef84edd0/dense_reference.npz", "window_s": [ 42355.508196607, 42375.474846357 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "63fee00a4aac01875997eb12651a7434ac0f53309082a8ebbde5d2ea95407202", "mano_sha256": "68f92333cca129942eb1da69624ed76f7a3490ea6f9301ce85cec944617443a7", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "7f75bf753835f1367a1963c5", "sample_id": "3831c7445959a2616fbef714", "file": "20_seconds/HOT3D/3831c7445959a2616fbef714_MANO_GT_20.00s.mp4", "recording_id": "P0010_fd1891a8", "source_first_frame": "158", "source_last_frame": "757", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "7f75bf753835f1367a1963c5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/3831c7445959a2616fbef714_MANO_GT_20.00s.mp4", "bytes": 6676306, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.06282745442944239, "p90": 0.3980946857287446, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "camera_rotation": { "median": 16.39542360139496, "p90": 46.35563283142818, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.09134158623974642, "p90": 0.29423863616501916, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "cmh_m_s": { "median": 0.09484309936969999, "p90": 0.2953768622124702, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.11626124881271543, "p90": 0.34276171672492883, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.11556695715533388, "p90": 0.34288519543862767, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "wms": { "median": 0.6790613370249341, "p90": 0.9406656618642195, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5284976633296296, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.198353742555198, "p90": 0.6532983176445764, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "cmh_m_s": { "median": 0.1370550751008412, "p90": 0.4255530004480347, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.2009145280391207, "p90": 0.5300602988658054, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.2010132597214382, "p90": 0.5306787566129125, "valid_pairs": 599, "valid_duration_s": 19.966641325001547, "coverage": 1.0 }, "wms": { "median": 0.6217115103989137, "p90": 0.9432341285156947, "valid_pairs": 598, "valid_duration_s": 19.933317647002696, "coverage": 0.9983310323726242 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6358581470790223, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.3980946857287446, "camera_angular_speed_p90_deg_s": 46.35563283142818, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6532983176445764, "hand_camera_speed_p90_m_s": 0.5306787566129125, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.4255530004480347, "rhcm_p90_m_s": 0.5306787566129125, "rham_p90_deg_s": null, "wms_p90": 0.9432341285156947, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "437a9c8d7433efce604af68d5ea83b9a7497913e9dba8926590660b124e2566e", "metadata_sha256": "4cc091415653143b1a701bb697730daa5fd7c16916762dc7f58057650a34528a", "source_file": "long20_20260918/local/train/previews/7f75bf753835f1367a1963c5/dense_reference.npz", "window_s": [ 56830.308214558, 56850.274855883 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "65eb4805da251f9e06cf162f7740e243b7e69a708a24ae5268faee112c1f5a79", "mano_sha256": "b4512722406b4ca5677a1c606f13f326cbcaf9ef0de012fa3d583d90814172e2", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "88f6a955e8947ce7f8fe9119", "sample_id": "eecfee28bef2df4c54f7a673", "file": "20_seconds/HOT3D/eecfee28bef2df4c54f7a673_MANO_GT_20.00s.mp4", "recording_id": "P0001_550ea2ac", "source_first_frame": "507", "source_last_frame": "1106", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "valid mesh not rasterized in 30 frames; per-hand official MANO coverage: left absent 0/600, right absent 13/600; absence alone does not establish visible-hand label failure", "id": "88f6a955e8947ce7f8fe9119", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/eecfee28bef2df4c54f7a673_MANO_GT_20.00s.mp4", "bytes": 5847098, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.20934845607566066, "p90": 0.5275251274368248, "valid_pairs": 599, "valid_duration_s": 19.96674725000048, "coverage": 1.0 }, "camera_rotation": { "median": 18.78839875769413, "p90": 61.70581459219067, "valid_pairs": 599, "valid_duration_s": 19.96674725000048, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.2409882449817725, "p90": 0.9101939811210045, "valid_pairs": 599, "valid_duration_s": 19.96674725000048, "coverage": 1.0 }, "cmh_m_s": { "median": 0.2059045809586713, "p90": 0.6566117280291469, "valid_pairs": 599, "valid_duration_s": 19.96674725000048, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.253781711787067, "p90": 0.6950303973116214, "valid_pairs": 599, "valid_duration_s": 19.96674725000048, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.25373793416314155, "p90": 0.6949059358283681, "valid_pairs": 599, "valid_duration_s": 19.96674725000048, "coverage": 1.0 }, "wms": { "median": 0.5450889436589702, "p90": 0.8858225927367482, "valid_pairs": 599, "valid_duration_s": 19.96674725000048, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6179079960469233, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.26113297017665255, "p90": 0.7476650873321905, "valid_pairs": 585, "valid_duration_s": 19.500085882995336, "coverage": 0.9766280726068122 }, "cmh_m_s": { "median": 0.21941070310637806, "p90": 0.7039109510968151, "valid_pairs": 585, "valid_duration_s": 19.500085882995336, "coverage": 0.9766280726068122 }, "rhcm_continuous_estimate_m_s": { "median": 0.21424424972725134, "p90": 0.581317480010988, "valid_pairs": 585, "valid_duration_s": 19.500085882995336, "coverage": 0.9766280726068122 }, "rhcm_discrete_m_s": { "median": 0.21479467225658827, "p90": 0.5807833147437994, "valid_pairs": 585, "valid_duration_s": 19.500085882995336, "coverage": 0.9766280726068122 }, "wms": { "median": 0.506066639844712, "p90": 0.8840645580660449, "valid_pairs": 585, "valid_duration_s": 19.500085882995336, "coverage": 0.9766280726068122 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5558177519956939, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.5275251274368248, "camera_angular_speed_p90_deg_s": 61.70581459219067, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.9101939811210045, "hand_camera_speed_p90_m_s": 0.6949059358283681, "hand_world_valid_fraction": 0.9891666666666666, "hand_camera_valid_fraction": 0.9891666666666666, "cmh_p90_m_s": 0.7039109510968151, "rhcm_p90_m_s": 0.6949059358283681, "rham_p90_deg_s": null, "wms_p90": 0.8858225927367482, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "ecf6b78dd5da59700646f367c32e284403149f106adbb1484d5f96cf4419b94f", "metadata_sha256": "2083b0411594ed76d9dff5cf0b8398798466ad8f06df525b22b3a066e0c778b9", "source_file": "long20_20260918/local/train/previews/88f6a955e8947ce7f8fe9119/dense_reference.npz", "window_s": [ 64427.074858317, 64447.041605567 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a0806f4e1ab2769e1ca56a42cda98bcaccdf3fde0dcc8e6b155c5eef6b21bfe7", "mano_sha256": "cf6314b7cacaf51d62f3ca9d4e9a77a2488f579b6652690ffe29224bf1f055f9", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "8c83f80145f1c87c26d946aa", "sample_id": "1e0c465c910a26b228bc7726", "file": "20_seconds/HOT3D/1e0c465c910a26b228bc7726_MANO_GT_20.00s.mp4", "recording_id": "P0014_9b7a0725", "source_first_frame": "3194", "source_last_frame": "3793", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 14/600 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 5 frames; per-hand official MANO coverage: left absent 19/600, right absent 18/600; absence alone does not establish visible-hand label failure", "id": "8c83f80145f1c87c26d946aa", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/1e0c465c910a26b228bc7726_MANO_GT_20.00s.mp4", "bytes": 4902792, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0609435981002993, "p90": 0.2111754885693853, "valid_pairs": 599, "valid_duration_s": 19.96666944500612, "coverage": 1.0 }, "camera_rotation": { "median": 11.423692030896266, "p90": 45.340096782767795, "valid_pairs": 599, "valid_duration_s": 19.96666944500612, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03240683381473103, "p90": 0.5033824408039531, "valid_pairs": 578, "valid_duration_s": 19.266688999006874, "coverage": 0.9649425534925997 }, "cmh_m_s": { "median": 0.07702712240402756, "p90": 0.36972394631526206, "valid_pairs": 578, "valid_duration_s": 19.266688999006874, "coverage": 0.9649425534925997 }, "rhcm_continuous_estimate_m_s": { "median": 0.08902493259981317, "p90": 0.5631497964094476, "valid_pairs": 578, "valid_duration_s": 19.266688999006874, "coverage": 0.9649425534925997 }, "rhcm_discrete_m_s": { "median": 0.0889797215757155, "p90": 0.5614046751186562, "valid_pairs": 578, "valid_duration_s": 19.266688999006874, "coverage": 0.9649425534925997 }, "wms": { "median": 0.8023535008424626, "p90": 0.9814811912053681, "valid_pairs": 544, "valid_duration_s": 18.133405215019593, "coverage": 0.9081837742125267 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.16744398940209074, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.008582516651731157, "p90": 0.04164393938439361, "valid_pairs": 580, "valid_duration_s": 19.33333236600447, "coverage": 0.9682802842634302 }, "cmh_m_s": { "median": 0.06188301445668784, "p90": 0.3702032820681338, "valid_pairs": 580, "valid_duration_s": 19.33333236600447, "coverage": 0.9682802842634302 }, "rhcm_continuous_estimate_m_s": { "median": 0.0633198430669577, "p90": 0.4177774754875263, "valid_pairs": 580, "valid_duration_s": 19.33333236600447, "coverage": 0.9682802842634302 }, "rhcm_discrete_m_s": { "median": 0.06325721861714725, "p90": 0.4191173487219845, "valid_pairs": 580, "valid_duration_s": 19.33333236600447, "coverage": 0.9682802842634302 }, "wms": { "median": 0.9174516350628186, "p90": 0.9905508855983366, "valid_pairs": 504, "valid_duration_s": 16.799948215993936, "coverage": 0.8413996266260514 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": -0.2358232234689287, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.2111754885693853, "camera_angular_speed_p90_deg_s": 45.340096782767795, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5033824408039531, "hand_camera_speed_p90_m_s": 0.5614046751186562, "hand_world_valid_fraction": 0.9691666666666666, "hand_camera_valid_fraction": 0.9691666666666666, "cmh_p90_m_s": 0.3702032820681338, "rhcm_p90_m_s": 0.5614046751186562, "rham_p90_deg_s": null, "wms_p90": 0.9905508855983366, "gt_coverage": 0.9682802842634302, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "bb9592f88b64bdd7c2ba7791e65aab478f4d4adf305254921b750c6264f55787", "metadata_sha256": "4014633d86c74dfe2890b26d0631db5aafa90d700b8058914b80479426f65374", "source_file": "long20_20260918/local/train/previews/8c83f80145f1c87c26d946aa/dense_reference.npz", "window_s": [ 43820.474871497, 43840.441540942 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "40c5886ca406b22545a31e9f63e09b598cbb82cc704aed9c4d4fda1f98545564", "mano_sha256": "b12116de12bdf9c0eaae151f4ee4d04a6489c61af956af942848ad56f6651558", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "9b325065536f771030ea57f1", "sample_id": "91f3cf6b8c418266063851e2", "file": "20_seconds/HOT3D/91f3cf6b8c418266063851e2_MANO_GT_20.00s.mp4", "recording_id": "P0012_476bae57", "source_first_frame": "1", "source_last_frame": "600", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "9b325065536f771030ea57f1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/91f3cf6b8c418266063851e2_MANO_GT_20.00s.mp4", "bytes": 7401052, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.027205574337855438, "p90": 0.1426473213642222, "valid_pairs": 599, "valid_duration_s": 19.966676509997342, "coverage": 1.0 }, "camera_rotation": { "median": 7.949507163682806, "p90": 33.10156480952142, "valid_pairs": 599, "valid_duration_s": 19.966676509997342, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03252604246415296, "p90": 0.18484609699779986, "valid_pairs": 599, "valid_duration_s": 19.966676509997342, "coverage": 1.0 }, "cmh_m_s": { "median": 0.052449787300105626, "p90": 0.26511365718086044, "valid_pairs": 599, "valid_duration_s": 19.966676509997342, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05149018801818112, "p90": 0.31726430141212036, "valid_pairs": 599, "valid_duration_s": 19.966676509997342, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.05152883461881876, "p90": 0.316973724601681, "valid_pairs": 599, "valid_duration_s": 19.966676509997342, "coverage": 1.0 }, "wms": { "median": 0.6701142209499592, "p90": 0.9616932811182397, "valid_pairs": 581, "valid_duration_s": 19.366730955014646, "coverage": 0.9699526581360547 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6580939660324139, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.13899776753045306, "p90": 0.4371900116874313, "valid_pairs": 591, "valid_duration_s": 19.699990232991695, "coverage": 0.9866434317763341 }, "cmh_m_s": { "median": 0.04330014532158016, "p90": 0.2778890725495481, "valid_pairs": 591, "valid_duration_s": 19.699990232991695, "coverage": 0.9866434317763341 }, "rhcm_continuous_estimate_m_s": { "median": 0.15769019098435483, "p90": 0.5062444567459947, "valid_pairs": 591, "valid_duration_s": 19.699990232991695, "coverage": 0.9866434317763341 }, "rhcm_discrete_m_s": { "median": 0.15771194396018587, "p90": 0.5073100730825235, "valid_pairs": 591, "valid_duration_s": 19.699990232991695, "coverage": 0.9866434317763341 }, "wms": { "median": 0.8368583553335329, "p90": 0.9716805526695, "valid_pairs": 589, "valid_duration_s": 19.63333162699564, "coverage": 0.9833049389649402 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6363896008721736, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1426473213642222, "camera_angular_speed_p90_deg_s": 33.10156480952142, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4371900116874313, "hand_camera_speed_p90_m_s": 0.5073100730825235, "hand_world_valid_fraction": 0.9941666666666666, "hand_camera_valid_fraction": 0.9941666666666666, "cmh_p90_m_s": 0.2778890725495481, "rhcm_p90_m_s": 0.5073100730825235, "rham_p90_deg_s": null, "wms_p90": 0.9716805526695, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "710e0d5b37f7a91a5581d1521db325af2cb1ffeb04a7ce21168afc909c41c363", "metadata_sha256": "48712f30bee3f0279a713bdbb21116139a9bd0c1d9a76f1abd542d7ab6faf637", "source_file": "long20_20260918/local/train/previews/9b325065536f771030ea57f1/dense_reference.npz", "window_s": [ 58673.0415193, 58693.00819581 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4bbfaa1c9720d8da573d181d8f0f5e4ec4219efbde35a53f4f723b3ac0de793b", "mano_sha256": "e916cb7f344418a0ac2ef5447126a5ea20fa650e3c27eeae0b96a71ea270228e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "b59242f998c93b9865429e72", "sample_id": "eac62f2d1892f715169b0328", "file": "20_seconds/HOT3D/eac62f2d1892f715169b0328_MANO_GT_20.00s.mp4", "recording_id": "P0014_1fdca00e", "source_first_frame": "2675", "source_last_frame": "3274", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "b59242f998c93b9865429e72", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/eac62f2d1892f715169b0328_MANO_GT_20.00s.mp4", "bytes": 5341868, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0504037886081822, "p90": 0.2562978526784224, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "camera_rotation": { "median": 11.873622127228524, "p90": 41.27951581500821, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.025245549360535986, "p90": 0.17536961334242504, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "cmh_m_s": { "median": 0.08638827319077755, "p90": 0.3653665491990404, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08593680594316386, "p90": 0.3567241306473014, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0858784893880472, "p90": 0.3566492017911582, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "wms": { "median": 0.7827954782220465, "p90": 0.9777302353399712, "valid_pairs": 565, "valid_duration_s": 18.83341145399754, "coverage": 0.9432438270477499 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5245039949525686, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.044438547450195376, "p90": 0.38867503265336395, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "cmh_m_s": { "median": 0.09603060559868772, "p90": 0.38968749059320684, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.11929892044165956, "p90": 0.4684874538890708, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.1192656296545615, "p90": 0.46794002432275406, "valid_pairs": 599, "valid_duration_s": 19.966641619001166, "coverage": 1.0 }, "wms": { "median": 0.7636934684354316, "p90": 0.9606652824270024, "valid_pairs": 569, "valid_duration_s": 18.96662266500789, "coverage": 0.9499155154344228 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5489962646774724, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.2562978526784224, "camera_angular_speed_p90_deg_s": 41.27951581500821, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.38867503265336395, "hand_camera_speed_p90_m_s": 0.46794002432275406, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.38968749059320684, "rhcm_p90_m_s": 0.46794002432275406, "rham_p90_deg_s": null, "wms_p90": 0.9777302353399712, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "832d55271de57b42871a7c73360304c7b12599d1ebe352e96fc29d93e99021b3", "metadata_sha256": "a099b0145abf4c4de2c70a2b9df0e7cea9f3c8efc6c8c06ada15a0aa11e6220e", "source_file": "long20_20260918/local/train/previews/b59242f998c93b9865429e72/dense_reference.npz", "window_s": [ 43161.174872177, 43181.141513796 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4e1060b40dd0cf0af7470aee3c01eb9b42823ff733b8641c1e0fbfe0d4b03acf", "mano_sha256": "75ce5a6ea72313381fd0a42c182915e0fc484ea8d8f371fb4d1bbca6b2141941", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "c8bc0903a5640f878be46a6d", "sample_id": "e4be7983ddfd9a427c50012c", "file": "20_seconds/HOT3D/e4be7983ddfd9a427c50012c_MANO_GT_20.00s.mp4", "recording_id": "P0014_84ea2dcc", "source_first_frame": "1700", "source_last_frame": "2299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "c8bc0903a5640f878be46a6d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/e4be7983ddfd9a427c50012c_MANO_GT_20.00s.mp4", "bytes": 4735084, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05783484273118335, "p90": 0.3330542611429599, "valid_pairs": 599, "valid_duration_s": 19.966661229998863, "coverage": 1.0 }, "camera_rotation": { "median": 9.916508576380618, "p90": 48.304241963139404, "valid_pairs": 599, "valid_duration_s": 19.966661229998863, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.015985517366955887, "p90": 0.35065284043586825, "valid_pairs": 428, "valid_duration_s": 14.266627641991363, "coverage": 0.7145224470757535 }, "cmh_m_s": { "median": 0.061020417237091125, "p90": 0.35232005267156896, "valid_pairs": 428, "valid_duration_s": 14.266627641991363, "coverage": 0.7145224470757535 }, "rhcm_continuous_estimate_m_s": { "median": 0.07169547879319671, "p90": 0.5246883519314924, "valid_pairs": 428, "valid_duration_s": 14.266627641991363, "coverage": 0.7145224470757535 }, "rhcm_discrete_m_s": { "median": 0.0715452147516753, "p90": 0.5247292092151619, "valid_pairs": 428, "valid_duration_s": 14.266627641991363, "coverage": 0.7145224470757535 }, "wms": { "median": 0.7512175091409087, "p90": 0.9731027356310011, "valid_pairs": 397, "valid_duration_s": 13.234950713987928, "coverage": 0.6628524700014997 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.37959925779751125, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1228589562268579, "p90": 0.6058771402105313, "valid_pairs": 599, "valid_duration_s": 19.966661229998863, "coverage": 1.0 }, "cmh_m_s": { "median": 0.07506028336799545, "p90": 0.38218593999322914, "valid_pairs": 599, "valid_duration_s": 19.966661229998863, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.10116959996219703, "p90": 0.3943396274654851, "valid_pairs": 599, "valid_duration_s": 19.966661229998863, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.10025080753287886, "p90": 0.39455226997078663, "valid_pairs": 599, "valid_duration_s": 19.966661229998863, "coverage": 1.0 }, "wms": { "median": 0.4019366492381577, "p90": 0.8794782262820925, "valid_pairs": 571, "valid_duration_s": 19.033406410999305, "coverage": 0.9532593452531066 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.8032415787739877, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.3330542611429599, "camera_angular_speed_p90_deg_s": 48.304241963139404, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6058771402105313, "hand_camera_speed_p90_m_s": 0.5247292092151619, "hand_world_valid_fraction": 0.86, "hand_camera_valid_fraction": 0.86, "cmh_p90_m_s": 0.38218593999322914, "rhcm_p90_m_s": 0.5247292092151619, "rham_p90_deg_s": null, "wms_p90": 0.9731027356310011, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "4f6912a023ecde087075ad8cef862b4239e1377e631a0e2ce061c6d10d9873a1", "metadata_sha256": "d51d3302ff989355f5f04c5e251dea9aecebc69551b8689f7b7455d61a5f4902", "source_file": "long20_20260918/local/train/previews/c8bc0903a5640f878be46a6d/dense_reference.npz", "window_s": [ 44801.674862312, 44821.641523542 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "1495e5648979d27421c6e7d614a67cbfdfa61a548cd513be6c520633a5ed1679", "mano_sha256": "c85c5960269565bd33474784740266ec0a93cf8eef3e1bb6e270b7092be64c34", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "官方 JSON 在 0.60–2.70 秒缺右手 MANO（63 帧），抽查 RGB 中右手可见；10.833–10.967 秒另缺 4 帧。逐帧源记录、解码 mesh 与渲染有效性一致，非渲染器漏显。当前原 T=8 训练窗口约 9.50–9.73 秒，不与这两段缺失重叠；应按区间审核，不整片自动排除。", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "c937aece0cc2a1cb17879859", "sample_id": "05f6c0bba9ea94097e30150a", "file": "20_seconds/HOT3D/05f6c0bba9ea94097e30150a_MANO_GT_20.00s.mp4", "recording_id": "P0010_160e551c", "source_first_frame": "1040", "source_last_frame": "1639", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 0/600, right absent 67/600; absence alone does not establish visible-hand label failure", "id": "c937aece0cc2a1cb17879859", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/05f6c0bba9ea94097e30150a_MANO_GT_20.00s.mp4", "bytes": 6656228, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03906251546489553, "p90": 0.23738729460734867, "valid_pairs": 599, "valid_duration_s": 19.96667110699491, "coverage": 1.0 }, "camera_rotation": { "median": 10.520450995076176, "p90": 30.112775961443074, "valid_pairs": 599, "valid_duration_s": 19.96667110699491, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.11044747057681535, "p90": 0.29550173891469916, "valid_pairs": 599, "valid_duration_s": 19.96667110699491, "coverage": 1.0 }, "cmh_m_s": { "median": 0.05949064070250612, "p90": 0.23290141022637642, "valid_pairs": 599, "valid_duration_s": 19.96667110699491, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.11391957485126865, "p90": 0.3140006633890668, "valid_pairs": 599, "valid_duration_s": 19.96667110699491, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.11385083334927644, "p90": 0.31422883662768203, "valid_pairs": 599, "valid_duration_s": 19.96667110699491, "coverage": 1.0 }, "wms": { "median": 0.6926438029171701, "p90": 0.9437856404943279, "valid_pairs": 599, "valid_duration_s": 19.96667110699491, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.46838057855623355, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1332609251490209, "p90": 0.45777300319541264, "valid_pairs": 530, "valid_duration_s": 17.668729953991715, "coverage": 0.8849111531567143 }, "cmh_m_s": { "median": 0.06240123945735455, "p90": 0.2978888519532048, "valid_pairs": 530, "valid_duration_s": 17.668729953991715, "coverage": 0.8849111531567143 }, "rhcm_continuous_estimate_m_s": { "median": 0.1371677528400491, "p90": 0.3816826651319021, "valid_pairs": 530, "valid_duration_s": 17.668729953991715, "coverage": 0.8849111531567143 }, "rhcm_discrete_m_s": { "median": 0.13718265951951042, "p90": 0.3813616661913427, "valid_pairs": 530, "valid_duration_s": 17.668729953991715, "coverage": 0.8849111531567143 }, "wms": { "median": 0.6589352616316433, "p90": 0.9397152880657591, "valid_pairs": 530, "valid_duration_s": 17.668729953991715, "coverage": 0.8849111531567143 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.42468509630836904, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "low", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.23738729460734867, "camera_angular_speed_p90_deg_s": 30.112775961443074, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.45777300319541264, "hand_camera_speed_p90_m_s": 0.3813616661913427, "hand_world_valid_fraction": 0.9441666666666667, "hand_camera_valid_fraction": 0.9441666666666667, "cmh_p90_m_s": 0.2978888519532048, "rhcm_p90_m_s": 0.3813616661913427, "rham_p90_deg_s": null, "wms_p90": 0.9437856404943279, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "cea910970e0fa00860ea1059aa04c0578a56d2488a91c16debe2a3374919cb89", "metadata_sha256": "ee9ad3efeed5cce639536c6b37633a7325700a95357eb78a498a5f510161c14f", "source_file": "long20_20260918/local/train/previews/c937aece0cc2a1cb17879859/dense_reference.npz", "window_s": [ 57197.674870055, 57217.641541162 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "20c0bc7ca7486a2ebb9345223d46a90bba5679341a210699d4d2c804dd30e93e", "mano_sha256": "7c3c0bc114016abf0dceb39482a7434e6267f024b6b33b8ebdaa846073dcaf7b", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "c9484c78ad9264add30a994d", "sample_id": "ae460103bc81297ce08d949f", "file": "20_seconds/HOT3D/ae460103bc81297ce08d949f_MANO_GT_20.00s.mp4", "recording_id": "P0001_b2bcbe28", "source_first_frame": "1234", "source_last_frame": "1833", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "valid mesh not rasterized in 12 frames; per-hand official MANO coverage: left absent 82/600, right absent 0/600; absence alone does not establish visible-hand label failure", "id": "c9484c78ad9264add30a994d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/ae460103bc81297ce08d949f_MANO_GT_20.00s.mp4", "bytes": 4549777, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.2231149476546781, "p90": 0.5076921619855994, "valid_pairs": 599, "valid_duration_s": 19.96666938999988, "coverage": 1.0 }, "camera_rotation": { "median": 21.60875364226026, "p90": 85.38440581055549, "valid_pairs": 599, "valid_duration_s": 19.96666938999988, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.33550690518705845, "p90": 0.6919702356918757, "valid_pairs": 417, "valid_duration_s": 13.899920833006036, "coverage": 0.696156207202373 }, "cmh_m_s": { "median": 0.28299620602877323, "p90": 0.5792527710864828, "valid_pairs": 417, "valid_duration_s": 13.899920833006036, "coverage": 0.696156207202373 }, "rhcm_continuous_estimate_m_s": { "median": 0.16033603726191226, "p90": 0.4618331189101942, "valid_pairs": 417, "valid_duration_s": 13.899920833006036, "coverage": 0.696156207202373 }, "rhcm_discrete_m_s": { "median": 0.1605842367306132, "p90": 0.4626164650474489, "valid_pairs": 417, "valid_duration_s": 13.899920833006036, "coverage": 0.696156207202373 }, "wms": { "median": 0.3456336973483318, "p90": 0.834106901692778, "valid_pairs": 416, "valid_duration_s": 13.866593352009659, "coverage": 0.6944870514535896 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6825057988206417, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.394927181049304, "p90": 0.9837098630437042, "valid_pairs": 599, "valid_duration_s": 19.96666938999988, "coverage": 1.0 }, "cmh_m_s": { "median": 0.23143894246705513, "p90": 0.621882803365938, "valid_pairs": 599, "valid_duration_s": 19.96666938999988, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.29169938533615475, "p90": 0.7478741387725721, "valid_pairs": 599, "valid_duration_s": 19.96666938999988, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.2894994503710529, "p90": 0.744958835499317, "valid_pairs": 599, "valid_duration_s": 19.96666938999988, "coverage": 1.0 }, "wms": { "median": 0.49619385968254826, "p90": 0.8968241767595875, "valid_pairs": 597, "valid_duration_s": 19.9000147340048, "coverage": 0.996661703827857 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5852194571777934, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.5076921619855994, "camera_angular_speed_p90_deg_s": 85.38440581055549, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.9837098630437042, "hand_camera_speed_p90_m_s": 0.744958835499317, "hand_world_valid_fraction": 0.8508333333333333, "hand_camera_valid_fraction": 0.8508333333333333, "cmh_p90_m_s": 0.621882803365938, "rhcm_p90_m_s": 0.744958835499317, "rham_p90_deg_s": null, "wms_p90": 0.8968241767595875, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "55ea9e8eb263eceb46e50cf9749ed2376d928dedc6fd055b6badf530156439bf", "metadata_sha256": "7e1ab3cc85c10208e5524690d8f0f3321f2de4839f6c1489e26220303f38f99b", "source_file": "long20_20260918/local/train/previews/c9484c78ad9264add30a994d/dense_reference.npz", "window_s": [ 54744.241530236, 54764.208199626 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "bdf9fbb745515c1269d090090b5007761d7abc763dfb4bf8b03635048f2feb57", "mano_sha256": "2d83cd8a0735e9d71bf19b051829d6b650b32b018f3bec82ca64bfe38f790a8c", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "caccc353afc8079786d16c15", "sample_id": "1a12f8686d876c9b3e59e6d2", "file": "20_seconds/HOT3D/1a12f8686d876c9b3e59e6d2_MANO_GT_20.00s.mp4", "recording_id": "P0012_130a66e1", "source_first_frame": "126", "source_last_frame": "725", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 29/600, right absent 0/600; absence alone does not establish visible-hand label failure", "id": "caccc353afc8079786d16c15", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/1a12f8686d876c9b3e59e6d2_MANO_GT_20.00s.mp4", "bytes": 5818506, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.027922005084674375, "p90": 0.21337328841364647, "valid_pairs": 599, "valid_duration_s": 19.96663654499571, "coverage": 1.0 }, "camera_rotation": { "median": 7.166637852155836, "p90": 39.09159803694417, "valid_pairs": 599, "valid_duration_s": 19.96663654499571, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06472954308486902, "p90": 0.4082973155429804, "valid_pairs": 517, "valid_duration_s": 17.2333185989919, "coverage": 0.8631057394246571 }, "cmh_m_s": { "median": 0.0399501073264227, "p90": 0.26505105347390845, "valid_pairs": 517, "valid_duration_s": 17.2333185989919, "coverage": 0.8631057394246571 }, "rhcm_continuous_estimate_m_s": { "median": 0.07322658548928987, "p90": 0.46800821490366734, "valid_pairs": 517, "valid_duration_s": 17.2333185989919, "coverage": 0.8631057394246571 }, "rhcm_discrete_m_s": { "median": 0.07307091399717446, "p90": 0.4682623141079785, "valid_pairs": 517, "valid_duration_s": 17.2333185989919, "coverage": 0.8631057394246571 }, "wms": { "median": 0.7790358825142198, "p90": 0.9422759214230151, "valid_pairs": 501, "valid_duration_s": 16.700031369007775, "coverage": 0.8363968228386142 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5611783538552464, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07133918433048624, "p90": 0.4150434738631795, "valid_pairs": 599, "valid_duration_s": 19.96663654499571, "coverage": 1.0 }, "cmh_m_s": { "median": 0.045468956936998955, "p90": 0.26589811531817725, "valid_pairs": 599, "valid_duration_s": 19.96663654499571, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.07338190125539568, "p90": 0.32236846910222816, "valid_pairs": 599, "valid_duration_s": 19.96663654499571, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0733723015698497, "p90": 0.32027674147747875, "valid_pairs": 599, "valid_duration_s": 19.96663654499571, "coverage": 1.0 }, "wms": { "median": 0.7110273877423661, "p90": 0.930538664953328, "valid_pairs": 565, "valid_duration_s": 18.83335515399085, "coverage": 0.943241247044741 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6720665434587187, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.21337328841364647, "camera_angular_speed_p90_deg_s": 39.09159803694417, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4150434738631795, "hand_camera_speed_p90_m_s": 0.4682623141079785, "hand_world_valid_fraction": 0.9333333333333333, "hand_camera_valid_fraction": 0.9333333333333333, "cmh_p90_m_s": 0.26589811531817725, "rhcm_p90_m_s": 0.4682623141079785, "rham_p90_deg_s": null, "wms_p90": 0.9422759214230151, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "64f69b4146af3d0201e007bc5bec14ecfab215743b7059d16465bde7aced514a", "metadata_sha256": "f5c9defd2e42d107fb6a26c26dce6715fa6f73087e0b0f2c62d029e31489e612", "source_file": "long20_20260918/local/train/previews/caccc353afc8079786d16c15/dense_reference.npz", "window_s": [ 46175.208216391, 46195.174852936 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "970172068ac66dcf3b05b5674c7952a092560bda3fe67ab81248f4982b7daa23", "mano_sha256": "f2a5fa0054850b680ce430fcdf1ef51f49993e586c642e9671ae4bb474113614", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "cef88fbe5569b10c0edb9255", "sample_id": "d2e2110bd40ff157f06c692c", "file": "20_seconds/HOT3D/d2e2110bd40ff157f06c692c_MANO_GT_20.00s.mp4", "recording_id": "P0003_e3a74169", "source_first_frame": "1614", "source_last_frame": "2213", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "cef88fbe5569b10c0edb9255", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/d2e2110bd40ff157f06c692c_MANO_GT_20.00s.mp4", "bytes": 6794704, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.038236812559955397, "p90": 0.20781899153085878, "valid_pairs": 599, "valid_duration_s": 19.966691441004514, "coverage": 1.0 }, "camera_rotation": { "median": 12.876257526132799, "p90": 53.402649842490334, "valid_pairs": 599, "valid_duration_s": 19.966691441004514, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.05078700646349167, "p90": 0.25647224751828096, "valid_pairs": 577, "valid_duration_s": 19.233305101006408, "coverage": 0.9632695110171338 }, "cmh_m_s": { "median": 0.08341987435957428, "p90": 0.4175711202103908, "valid_pairs": 577, "valid_duration_s": 19.233305101006408, "coverage": 0.9632695110171338 }, "rhcm_continuous_estimate_m_s": { "median": 0.11181339419095257, "p90": 0.4409909298713625, "valid_pairs": 577, "valid_duration_s": 19.233305101006408, "coverage": 0.9632695110171338 }, "rhcm_discrete_m_s": { "median": 0.11182171518479685, "p90": 0.44151480894635786, "valid_pairs": 577, "valid_duration_s": 19.233305101006408, "coverage": 0.9632695110171338 }, "wms": { "median": 0.802462242064085, "p90": 0.9719426370674583, "valid_pairs": 565, "valid_duration_s": 18.833298461016966, "coverage": 0.9432358143392771 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.38453499837084393, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.045258377711664126, "p90": 0.483808660993969, "valid_pairs": 599, "valid_duration_s": 19.966691441004514, "coverage": 1.0 }, "cmh_m_s": { "median": 0.08588596939342692, "p90": 0.4023657782864986, "valid_pairs": 599, "valid_duration_s": 19.966691441004514, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.10720946955409158, "p90": 0.4732663643873903, "valid_pairs": 599, "valid_duration_s": 19.966691441004514, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.10699533027020502, "p90": 0.4727790135246207, "valid_pairs": 599, "valid_duration_s": 19.966691441004514, "coverage": 1.0 }, "wms": { "median": 0.7574978823277881, "p90": 0.9591323327815955, "valid_pairs": 584, "valid_duration_s": 19.466761174000567, "coverage": 0.9749617873105773 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.5597016208731385, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.20781899153085878, "camera_angular_speed_p90_deg_s": 53.402649842490334, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.483808660993969, "hand_camera_speed_p90_m_s": 0.4727790135246207, "hand_world_valid_fraction": 0.9833333333333333, "hand_camera_valid_fraction": 0.9833333333333333, "cmh_p90_m_s": 0.4175711202103908, "rhcm_p90_m_s": 0.4727790135246207, "rham_p90_deg_s": null, "wms_p90": 0.9719426370674583, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "78b000b3862f02ac4e14ba1e62b30e1f40a65ca468cc5c5a86e91dd0f45d141f", "metadata_sha256": "3d48ae1c30b1408c8ee43f6d71b639052a21df3dc6b602807210106c265aac55", "source_file": "long20_20260918/local/train/previews/cef88fbe5569b10c0edb9255/dense_reference.npz", "window_s": [ 47068.908193423, 47088.874884864 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "28f15907a2d325a725f5ffaed6e5c571480d16cfcaed0fbc6a54972e9e785709", "mano_sha256": "00f63b0d76ff3453e89a541db8060d38e862257d999585dce7f128f5f4d8caf4", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "d130dcb7355197be146e15d9", "sample_id": "d9f0284cdf6285324fe277d2", "file": "20_seconds/HOT3D/d9f0284cdf6285324fe277d2_MANO_GT_20.00s.mp4", "recording_id": "P0009_02511c2f", "source_first_frame": "2775", "source_last_frame": "3374", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "valid mesh not rasterized in 8 frames; per-hand official MANO coverage: left absent 46/600, right absent 0/600; absence alone does not establish visible-hand label failure", "id": "d130dcb7355197be146e15d9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/d9f0284cdf6285324fe277d2_MANO_GT_20.00s.mp4", "bytes": 6370699, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08224242246058167, "p90": 0.3196620591308563, "valid_pairs": 599, "valid_duration_s": 19.968274605998886, "coverage": 1.0 }, "camera_rotation": { "median": 13.305579472057552, "p90": 48.09557924033074, "valid_pairs": 599, "valid_duration_s": 19.968274605998886, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.052246550710174575, "p90": 0.3986014622838074, "valid_pairs": 552, "valid_duration_s": 18.401607318999595, "coverage": 0.9215421803880526 }, "cmh_m_s": { "median": 0.12533471004542063, "p90": 0.4998958407137534, "valid_pairs": 552, "valid_duration_s": 18.401607318999595, "coverage": 0.9215421803880526 }, "rhcm_continuous_estimate_m_s": { "median": 0.11844011488207412, "p90": 0.43472920305847584, "valid_pairs": 552, "valid_duration_s": 18.401607318999595, "coverage": 0.9215421803880526 }, "rhcm_discrete_m_s": { "median": 0.11836609687070115, "p90": 0.4345065979145765, "valid_pairs": 552, "valid_duration_s": 18.401607318999595, "coverage": 0.9215421803880526 }, "wms": { "median": 0.6361855302757708, "p90": 0.9582433538263452, "valid_pairs": 542, "valid_duration_s": 18.068255777994636, "coverage": 0.9048481220589062 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6191680737151173, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.21467042986199425, "p90": 0.573132641362777, "valid_pairs": 599, "valid_duration_s": 19.968274605998886, "coverage": 1.0 }, "cmh_m_s": { "median": 0.12245302454246637, "p90": 0.4677607685068443, "valid_pairs": 599, "valid_duration_s": 19.968274605998886, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.17458075659753264, "p90": 0.572530734187252, "valid_pairs": 599, "valid_duration_s": 19.968274605998886, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.17495524664916437, "p90": 0.5699178547040181, "valid_pairs": 599, "valid_duration_s": 19.968274605998886, "coverage": 1.0 }, "wms": { "median": 0.6188604010302882, "p90": 0.9262710888996514, "valid_pairs": 599, "valid_duration_s": 19.968274605998886, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4943344258267681, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.3196620591308563, "camera_angular_speed_p90_deg_s": 48.09557924033074, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.573132641362777, "hand_camera_speed_p90_m_s": 0.5699178547040181, "hand_world_valid_fraction": 0.9616666666666667, "hand_camera_valid_fraction": 0.9616666666666667, "cmh_p90_m_s": 0.4998958407137534, "rhcm_p90_m_s": 0.5699178547040181, "rham_p90_deg_s": null, "wms_p90": 0.9582433538263452, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "b103b8787eb737bd23b9fd93cc7811f3471bb6c77c38be2b239dae9c39c0b253", "metadata_sha256": "316ec76377366c63fcd2c948a57cb3f4cffde24b33b09d6399e37c82e4aa7957", "source_file": "long20_20260918/local/benchmark/previews/d130dcb7355197be146e15d9/dense_reference.npz", "window_s": [ 40954.508270492, 40974.476545098 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5086e5bf549aa7fd301c3eee6925ed3b6363cd7e186d67d65493e70018f588ec", "mano_sha256": "fd9f18d68a05c8c9ae954a1a8d7dfced9183c07a8493d5335cc47d10d8a35533", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "dedc0e251c80019111291533", "sample_id": "5b80aa5d9c6368abb912e7d8", "file": "20_seconds/HOT3D/5b80aa5d9c6368abb912e7d8_MANO_GT_20.00s.mp4", "recording_id": "P0014_65c6a968", "source_first_frame": "466", "source_last_frame": "1065", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "dedc0e251c80019111291533", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/5b80aa5d9c6368abb912e7d8_MANO_GT_20.00s.mp4", "bytes": 4888177, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.029086227511277544, "p90": 0.1422184584870268, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "camera_rotation": { "median": 7.154484870780361, "p90": 40.38435223351941, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.003924015822338263, "p90": 0.02499338900165052, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "cmh_m_s": { "median": 0.044100739607460505, "p90": 0.2664384510122856, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.04605230594034495, "p90": 0.2747651998711184, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.046068915426278875, "p90": 0.2744347524593116, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "wms": { "median": 0.9332367603068072, "p90": 0.9850721027059937, "valid_pairs": 514, "valid_duration_s": 17.133268004974525, "coverage": 0.8580943677268428 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.2591996136258312, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.023549874803204734, "p90": 0.18532550712323237, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "cmh_m_s": { "median": 0.04785008805224406, "p90": 0.27232835669331007, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08012828833163127, "p90": 0.28107859100556815, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.08114251654291604, "p90": 0.28122881477759615, "valid_pairs": 599, "valid_duration_s": 19.966647782996006, "coverage": 1.0 }, "wms": { "median": 0.8284954593403807, "p90": 0.9716258305153694, "valid_pairs": 569, "valid_duration_s": 18.96660019900446, "coverage": 0.9499140970051465 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.19703200987152497, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1422184584870268, "camera_angular_speed_p90_deg_s": 40.38435223351941, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.18532550712323237, "hand_camera_speed_p90_m_s": 0.28122881477759615, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.27232835669331007, "rhcm_p90_m_s": 0.28122881477759615, "rham_p90_deg_s": null, "wms_p90": 0.9850721027059937, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "ab529a6da668a26e215c89b0642e8e1bef218390fa938ba3df30db0ec6b42920", "metadata_sha256": "a453af7da2a7b023057e61beee9519258d8d1d1328c8825ea13ba8c200f547e5", "source_file": "long20_20260918/local/train/previews/dedc0e251c80019111291533/dense_reference.npz", "window_s": [ 44237.541543288, 44257.508191071 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6713787c23d15d6db84c57c6e011740d7cc195ef180ee1df29d303663a067cb7", "mano_sha256": "5865d12b47ae0f4fb1a23831c8539b573f18ad5671d5ba7f8aa7ef925531dadd", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e46ca1e8d79b17a340aa73f8", "sample_id": "a663831621f88158c97734ec", "file": "20_seconds/HOT3D/a663831621f88158c97734ec_MANO_GT_20.00s.mp4", "recording_id": "P0011_dbf1ebb7", "source_first_frame": "2389", "source_last_frame": "2988", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 107/600, right absent 36/600; absence alone does not establish visible-hand label failure", "id": "e46ca1e8d79b17a340aa73f8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/a663831621f88158c97734ec_MANO_GT_20.00s.mp4", "bytes": 5203203, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05211536101600134, "p90": 0.22692895244770997, "valid_pairs": 599, "valid_duration_s": 19.96665310300159, "coverage": 1.0 }, "camera_rotation": { "median": 16.01365453837066, "p90": 42.8253968891991, "valid_pairs": 599, "valid_duration_s": 19.96665310300159, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0499625282660826, "p90": 0.10095025274239514, "valid_pairs": 395, "valid_duration_s": 13.16658460599865, "coverage": 0.6594287253890997 }, "cmh_m_s": { "median": 0.10770894216693216, "p90": 0.40955160169908117, "valid_pairs": 395, "valid_duration_s": 13.16658460599865, "coverage": 0.6594287253890997 }, "rhcm_continuous_estimate_m_s": { "median": 0.11635229509183471, "p90": 0.4126617893964453, "valid_pairs": 395, "valid_duration_s": 13.16658460599865, "coverage": 0.6594287253890997 }, "rhcm_discrete_m_s": { "median": 0.11658246140308823, "p90": 0.41071471524896896, "valid_pairs": 395, "valid_duration_s": 13.16658460599865, "coverage": 0.6594287253890997 }, "wms": { "median": 0.8179716468330668, "p90": 0.9702695825836657, "valid_pairs": 395, "valid_duration_s": 13.16658460599865, "coverage": 0.6594287253890997 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.2117021918731101, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.05689327809958828, "p90": 0.44940844874850744, "valid_pairs": 561, "valid_duration_s": 18.69998741299787, "coverage": 0.9365609407110247 }, "cmh_m_s": { "median": 0.09898251043790668, "p90": 0.2973745898216632, "valid_pairs": 561, "valid_duration_s": 18.69998741299787, "coverage": 0.9365609407110247 }, "rhcm_continuous_estimate_m_s": { "median": 0.10744347796378563, "p90": 0.3382253631690328, "valid_pairs": 561, "valid_duration_s": 18.69998741299787, "coverage": 0.9365609407110247 }, "rhcm_discrete_m_s": { "median": 0.1074495756982927, "p90": 0.33697297810520604, "valid_pairs": 561, "valid_duration_s": 18.69998741299787, "coverage": 0.9365609407110247 }, "wms": { "median": 0.6757252204085348, "p90": 0.9400868896273503, "valid_pairs": 557, "valid_duration_s": 18.568695761998242, "coverage": 0.9299853944578627 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.49482864320105263, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.22692895244770997, "camera_angular_speed_p90_deg_s": 42.8253968891991, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.44940844874850744, "hand_camera_speed_p90_m_s": 0.41071471524896896, "hand_world_valid_fraction": 0.8016666666666666, "hand_camera_valid_fraction": 0.8016666666666666, "cmh_p90_m_s": 0.40955160169908117, "rhcm_p90_m_s": 0.41071471524896896, "rham_p90_deg_s": null, "wms_p90": 0.9702695825836657, "gt_coverage": 0.9365609407110247, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "17cd6cf486fa40a7e2e90ba79bc103ded56cd220574d25e2eb81295aa3de426f", "metadata_sha256": "a9e9b473677e2aa98a24058e106a9a4ba253701657b8a8113e25fe42b95dcdbb", "source_file": "long20_20260918/local/train/previews/e46ca1e8d79b17a340aa73f8/dense_reference.npz", "window_s": [ 59814.641541409, 59834.608194512 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d2200e58a523b7ae1196bee0f5c95b62d0a0831d27fc1d756a45a04de4d87fed", "mano_sha256": "1e413c4ecffea402dd7fff089ab62e37f7df6e0ed85e40b886168394e5dc934e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e4b7414af19675b8a657bb60", "sample_id": "0d902764bd011510a3922a68", "file": "20_seconds/HOT3D/0d902764bd011510a3922a68_MANO_GT_20.00s.mp4", "recording_id": "P0012_c4353c31", "source_first_frame": "3005", "source_last_frame": "3604", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "e4b7414af19675b8a657bb60", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/0d902764bd011510a3922a68_MANO_GT_20.00s.mp4", "bytes": 5986694, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.061875248918405755, "p90": 0.196613226098151, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "camera_rotation": { "median": 14.16443585223838, "p90": 41.88375048666399, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.10943100250641109, "p90": 0.41517510705312916, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "cmh_m_s": { "median": 0.10219960788133503, "p90": 0.2642636909866436, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.17170973811220655, "p90": 0.44693419893383, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.17124497323917445, "p90": 0.4464353915635394, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "wms": { "median": 0.7626553583891827, "p90": 0.9528830440760636, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.32479237971870617, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.055540553934500955, "p90": 0.18945335596143226, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "cmh_m_s": { "median": 0.09847387628297771, "p90": 0.27061904065939907, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.10252301431544227, "p90": 0.3290501854726211, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.10271882174592434, "p90": 0.3294075485741769, "valid_pairs": 599, "valid_duration_s": 19.96664526600216, "coverage": 1.0 }, "wms": { "median": 0.7013335002456592, "p90": 0.9241761318276024, "valid_pairs": 595, "valid_duration_s": 19.833301625003514, "coverage": 0.9933216802711623 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4319275715936817, "hand_class": "low", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.196613226098151, "camera_angular_speed_p90_deg_s": 41.88375048666399, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.41517510705312916, "hand_camera_speed_p90_m_s": 0.4464353915635394, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.27061904065939907, "rhcm_p90_m_s": 0.4464353915635394, "rham_p90_deg_s": null, "wms_p90": 0.9528830440760636, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "4a514334c9482ae5115953c05e5a527248132a18270b5e571a7fe384116c2f1e", "metadata_sha256": "988cbf1645c79628286e1553a2ff5e1ed54f84724f282055baea9ba64313f3d0", "source_file": "long20_20260918/local/train/previews/e4b7414af19675b8a657bb60/dense_reference.npz", "window_s": [ 45557.174879727, 45577.141524993 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6d8dbd9a80b9f3714685de7f3ffdd6d82203915200f5e00d9e1339e76e7de092", "mano_sha256": "c2dbca81dc418e8be98c3e18dc56d330ebe5b0ff31cdf168b39f514be5f1ed1c", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "ea3320838670cebafcc7fbc5", "sample_id": "0a85f659684d2e41812d3d50", "file": "20_seconds/HOT3D/0a85f659684d2e41812d3d50_MANO_GT_20.00s.mp4", "recording_id": "P0001_a9d6c83d", "source_first_frame": "2423", "source_last_frame": "3022", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 30/600 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 14 frames; per-hand official MANO coverage: left absent 0/600, right absent 58/600; absence alone does not establish visible-hand label failure", "id": "ea3320838670cebafcc7fbc5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/0a85f659684d2e41812d3d50_MANO_GT_20.00s.mp4", "bytes": 6442160, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.3390034464047389, "p90": 0.9306394787466727, "valid_pairs": 599, "valid_duration_s": 19.96665695399861, "coverage": 1.0 }, "camera_rotation": { "median": 34.848881299288095, "p90": 81.59807541326508, "valid_pairs": 599, "valid_duration_s": 19.96665695399861, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.3711089803917024, "p90": 0.6727797893161601, "valid_pairs": 570, "valid_duration_s": 18.999751901996206, "coverage": 0.9515740139057798 }, "cmh_m_s": { "median": 0.30705667009487875, "p90": 0.7311148139896536, "valid_pairs": 570, "valid_duration_s": 18.999751901996206, "coverage": 0.9515740139057798 }, "rhcm_continuous_estimate_m_s": { "median": 0.24520045672220284, "p90": 0.6325047300768963, "valid_pairs": 570, "valid_duration_s": 18.999751901996206, "coverage": 0.9515740139057798 }, "rhcm_discrete_m_s": { "median": 0.24670946362859258, "p90": 0.6330429066604335, "valid_pairs": 570, "valid_duration_s": 18.999751901996206, "coverage": 0.9515740139057798 }, "wms": { "median": 0.397577154663571, "p90": 0.7913451943923558, "valid_pairs": 570, "valid_duration_s": 18.999751901996206, "coverage": 0.9515740139057798 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6164024415036836, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.3307194166955831, "p90": 0.8349065529337347, "valid_pairs": 539, "valid_duration_s": 17.96665714999108, "coverage": 0.8998330161821606 }, "cmh_m_s": { "median": 0.27893001247594723, "p90": 0.7089154985420872, "valid_pairs": 539, "valid_duration_s": 17.96665714999108, "coverage": 0.8998330161821606 }, "rhcm_continuous_estimate_m_s": { "median": 0.20050812060738896, "p90": 0.5341823850930104, "valid_pairs": 539, "valid_duration_s": 17.96665714999108, "coverage": 0.8998330161821606 }, "rhcm_discrete_m_s": { "median": 0.2000416720573411, "p90": 0.5304566241728661, "valid_pairs": 539, "valid_duration_s": 17.96665714999108, "coverage": 0.8998330161821606 }, "wms": { "median": 0.370528455449375, "p90": 0.6786424118289124, "valid_pairs": 539, "valid_duration_s": 17.96665714999108, "coverage": 0.8998330161821606 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.6263968874696436, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.9306394787466727, "camera_angular_speed_p90_deg_s": 81.59807541326508, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.8349065529337347, "hand_camera_speed_p90_m_s": 0.6330429066604335, "hand_world_valid_fraction": 0.9283333333333333, "hand_camera_valid_fraction": 0.9283333333333333, "cmh_p90_m_s": 0.7311148139896536, "rhcm_p90_m_s": 0.6330429066604335, "rham_p90_deg_s": null, "wms_p90": 0.7913451943923558, "gt_coverage": 0.9515740139057798, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "06852ac41950f820fabf5f3508f5c57b19746527d557995234a3681047d48c45", "metadata_sha256": "cb6d30bf07fd48b9a941b1096e12dcf14c87cf45bea9787ab9e7815e43f9d1dd", "source_file": "long20_20260918/local/train/previews/ea3320838670cebafcc7fbc5/dense_reference.npz", "window_s": [ 58840.774906086, 58860.74156304 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ece66d34d0bc8c84c7f544f4f4c8e1b50e0a9f0721e95bd630f50d005b7e5a25", "mano_sha256": "3783a8b61aace771f7ca2c6b0a137b296a31c67a8b6210e0f408f5816f8930ff", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "f32d7c7295b933b5d05ca5cd", "sample_id": "74b8bec89d5120f0909be79d", "file": "20_seconds/HOT3D/74b8bec89d5120f0909be79d_MANO_GT_20.00s.mp4", "recording_id": "P0002_2ea9af5b", "source_first_frame": "655", "source_last_frame": "1254", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "f32d7c7295b933b5d05ca5cd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/74b8bec89d5120f0909be79d_MANO_GT_20.00s.mp4", "bytes": 5952678, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0775244114538116, "p90": 0.18243825207383996, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "camera_rotation": { "median": 14.329696902056645, "p90": 47.149041423025004, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.037862940110866695, "p90": 0.48596918199742745, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "cmh_m_s": { "median": 0.09569588499880273, "p90": 0.3178397824338983, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.14701136860791647, "p90": 0.546609357510528, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.14657226979559132, "p90": 0.5465221522181232, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "wms": { "median": 0.8414042756503151, "p90": 0.9745668233990142, "valid_pairs": 587, "valid_duration_s": 19.56669129399961, "coverage": 0.9799682679375255 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.44890704127838477, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.2067359777525749, "p90": 0.5278409609074408, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "cmh_m_s": { "median": 0.103986325425466, "p90": 0.34339249611536915, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.19712927903338373, "p90": 0.4643124154721938, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.19724631357744954, "p90": 0.46331615453905994, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "wms": { "median": 0.6365985032698456, "p90": 0.9111431015822115, "valid_pairs": 599, "valid_duration_s": 19.966658037999878, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4875168759526746, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.18243825207383996, "camera_angular_speed_p90_deg_s": 47.149041423025004, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5278409609074408, "hand_camera_speed_p90_m_s": 0.5465221522181232, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.34339249611536915, "rhcm_p90_m_s": 0.5465221522181232, "rham_p90_deg_s": null, "wms_p90": 0.9745668233990142, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "ceec8385e1d0975ba13b00dbc4534a32fc971b1a81985c16cf2c20050fcb3c8e", "metadata_sha256": "c144a9a6f88b509fb58d874864a17f6447950cc628de0baa77c70d221cda5962", "source_file": "long20_20260918/local/train/previews/f32d7c7295b933b5d05ca5cd/dense_reference.npz", "window_s": [ 47877.841535815, 47897.808193853 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2693ce79085bdfbca0797d9381d2d3f475418189858e78aae4fbc0d95371a3e8", "mano_sha256": "c421eceef14dda60a9d05caf7954ba1bc0f17deb8a089a0ec2955b4803a79d53", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "官方 MANO 在操作区可见，场景云有孔洞和分层，不等于手标签错误。", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "f68bf6deb3a5c4642114799d", "sample_id": "fea61e05460db631c35b9c0c", "file": "20_seconds/HOT3D/fea61e05460db631c35b9c0c_MANO_GT_20.00s.mp4", "recording_id": "P0010_8ff3e5c4", "source_first_frame": "2233", "source_last_frame": "2832", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "valid mesh not rasterized in 18 frames; per-hand official MANO coverage: left absent 0/600, right absent 13/600; absence alone does not establish visible-hand label failure", "id": "f68bf6deb3a5c4642114799d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/fea61e05460db631c35b9c0c_MANO_GT_20.00s.mp4", "bytes": 5595639, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.09550101632038951, "p90": 0.290850105527277, "valid_pairs": 599, "valid_duration_s": 19.966674581002735, "coverage": 1.0 }, "camera_rotation": { "median": 15.876320276101081, "p90": 50.82739576625216, "valid_pairs": 599, "valid_duration_s": 19.966674581002735, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.062486952628536985, "p90": 0.2535389642683745, "valid_pairs": 581, "valid_duration_s": 19.36667395599943, "coverage": 0.9699498971363928 }, "cmh_m_s": { "median": 0.060705783860519835, "p90": 0.23267156445656006, "valid_pairs": 581, "valid_duration_s": 19.36667395599943, "coverage": 0.9699498971363928 }, "rhcm_continuous_estimate_m_s": { "median": 0.07823568531321634, "p90": 0.30059905248864305, "valid_pairs": 581, "valid_duration_s": 19.36667395599943, "coverage": 0.9699498971363928 }, "rhcm_discrete_m_s": { "median": 0.07855728850320617, "p90": 0.2992132840794854, "valid_pairs": 581, "valid_duration_s": 19.36667395599943, "coverage": 0.9699498971363928 }, "wms": { "median": 0.6065786813335432, "p90": 0.9086840997534945, "valid_pairs": 580, "valid_duration_s": 19.333312981994823, "coverage": 0.9682790643760717 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.4503366775721035, "hand_class": "low", "relative_class": "low", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.17758444959523914, "p90": 0.8083308317812259, "valid_pairs": 580, "valid_duration_s": 19.33339623199572, "coverage": 0.9682832338235459 }, "cmh_m_s": { "median": 0.11862354751244722, "p90": 0.4493389171561117, "valid_pairs": 580, "valid_duration_s": 19.33339623199572, "coverage": 0.9682832338235459 }, "rhcm_continuous_estimate_m_s": { "median": 0.15226060251300919, "p90": 0.5452554420922062, "valid_pairs": 580, "valid_duration_s": 19.33339623199572, "coverage": 0.9682832338235459 }, "rhcm_discrete_m_s": { "median": 0.15107000527989975, "p90": 0.5476806108589648, "valid_pairs": 580, "valid_duration_s": 19.33339623199572, "coverage": 0.9682832338235459 }, "wms": { "median": 0.5450510486186394, "p90": 0.8254418127980856, "valid_pairs": 579, "valid_duration_s": 19.30006866099575, "coverage": 0.9666140740010243 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.732434851840041, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.290850105527277, "camera_angular_speed_p90_deg_s": 50.82739576625216, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.8083308317812259, "hand_camera_speed_p90_m_s": 0.5476806108589648, "hand_world_valid_fraction": 0.9725, "hand_camera_valid_fraction": 0.9725, "cmh_p90_m_s": 0.4493389171561117, "rhcm_p90_m_s": 0.5476806108589648, "rham_p90_deg_s": null, "wms_p90": 0.9086840997534945, "gt_coverage": 0.9699498971363928, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "78993034e22ca5a65dd1b3cdcf6bb4e8cd091db7e57e15a71b02051ae1b07f50", "metadata_sha256": "2b4ad605eecfaf4fc20aabf7cfd29e9aa81dd46329a934c05462f36022c34b00", "source_file": "long20_20260918/local/train/previews/f68bf6deb3a5c4642114799d/dense_reference.npz", "window_s": [ 58769.441532929, 58789.40820751 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOT3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4970b85543c4dc314f4ca036d1604ad5e9e166786ee031c38c2ab46bd53cab70", "mano_sha256": "95d24d892f37dd35f4da0ce18deb0dcd40db5028339aea04e3a2257908d2222b", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native source MANO (official fitted annotation)", "scene_depth_source": "native Aria MPS semidense SLAM point map, RGB colorized with calibrated same-frame projections", "scene_camera_source": "official HOT3D headset trajectory and camera extrinsics", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source SLAM metres; no rescaling", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "fa99bdb8bf9fb6c73ec8bc88", "sample_id": "d2b9808e08e06502e0a10cdd", "file": "20_seconds/HOT3D/d2b9808e08e06502e0a10cdd_MANO_GT_20.00s.mp4", "recording_id": "P0009_ea91844d", "source_first_frame": "1275", "source_last_frame": "1874", "source_split": "external_heldout", "training_label_source": "missing", "quality_warnings": "per-hand official MANO coverage: left absent 0/600, right absent 4/600; absence alone does not establish visible-hand label failure", "id": "fa99bdb8bf9fb6c73ec8bc88", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOT3D/d2b9808e08e06502e0a10cdd_MANO_GT_20.00s.mp4", "bytes": 5885126, "width": 962, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.048079256748790405, "p90": 0.12735605212655002, "valid_pairs": 599, "valid_duration_s": 19.966658790996007, "coverage": 1.0 }, "camera_rotation": { "median": 10.61220951193858, "p90": 31.708368505855013, "valid_pairs": 599, "valid_duration_s": 19.966658790996007, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0875926668755459, "p90": 0.33644045919589555, "valid_pairs": 599, "valid_duration_s": 19.966658790996007, "coverage": 1.0 }, "cmh_m_s": { "median": 0.06674261461222143, "p90": 0.19551963635204206, "valid_pairs": 599, "valid_duration_s": 19.966658790996007, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.11549247021265487, "p90": 0.39904557100518384, "valid_pairs": 599, "valid_duration_s": 19.966658790996007, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.11543490391102948, "p90": 0.3990921718531256, "valid_pairs": 599, "valid_duration_s": 19.966658790996007, "coverage": 1.0 }, "wms": { "median": 0.7061325764141222, "p90": 0.9391050306093384, "valid_pairs": 588, "valid_duration_s": 19.5999825970066, "coverage": 0.9816355756950803 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.26969335737935574, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.10289946425669493, "p90": 0.44637223456106717, "valid_pairs": 594, "valid_duration_s": 19.799983366996457, "coverage": 0.9916523127006752 }, "cmh_m_s": { "median": 0.06965859687120188, "p90": 0.19577362472085205, "valid_pairs": 594, "valid_duration_s": 19.799983366996457, "coverage": 0.9916523127006752 }, "rhcm_continuous_estimate_m_s": { "median": 0.1287193360717337, "p90": 0.4310252283729368, "valid_pairs": 594, "valid_duration_s": 19.799983366996457, "coverage": 0.9916523127006752 }, "rhcm_discrete_m_s": { "median": 0.1285952732576875, "p90": 0.4310293617726092, "valid_pairs": 594, "valid_duration_s": 19.799983366996457, "coverage": 0.9916523127006752 }, "wms": { "median": 0.7335521328816159, "p90": 0.937111167448829, "valid_pairs": 586, "valid_duration_s": 19.533263466997596, "coverage": 0.9782940486670784 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": 0.24691505300571856, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.12735605212655002, "camera_angular_speed_p90_deg_s": 31.708368505855013, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.44637223456106717, "hand_camera_speed_p90_m_s": 0.4310293617726092, "hand_world_valid_fraction": 0.9966666666666667, "hand_camera_valid_fraction": 0.9966666666666667, "cmh_p90_m_s": 0.19577362472085205, "rhcm_p90_m_s": 0.4310293617726092, "rham_p90_deg_s": null, "wms_p90": 0.9391050306093384, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_reference_tracking", "camera_source": "official headset reference tracking + per-camera extrinsics, not independent mocap claim", "hand_source": "official HOT3D fitted MANO decoded by official MANOHandDataProvider, wrist+finger order explicitly mapped", "reference_sha256": "b21e3c1374b355b4ac99ff827f8da521368b6b99e6552042cb9ae7b491db6461", "metadata_sha256": "31926e3326d2f54248a004664e8ea20b1246fabd131200f8bdbcdb9228143637", "source_file": "long20_20260918/local/benchmark/previews/fa99bdb8bf9fb6c73ec8bc88/dense_reference.npz", "window_s": [ 41425.508195415, 41445.474854206 ], "timestamps_source": "source frame timeline", "frame_count": 600, "common_world_residual_max_m": 0.0 }, "camera_display_source": "official HOT3D headset trajectory and camera extrinsics", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方 SLAM", "kind": "estimated", "detail": "官方半稠密 SLAM 估计点图，由我们用 RGB 着色；不是独立测量 GT。" }, "camera": { "label": "官方参考跟踪", "kind": "estimated", "detail": "官方头显参考跟踪结合相机外参；不称独立 mocap GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "fcadfdd631e6283215e073a46de5f5fa393dc86372ad4064a36d8e77e034fa2d", "mano_sha256": "a3f864528a95caac877d2d24974b73286d20dd0cbf97d7d43444588d794f6d7c", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "006863bb291ad3c5d17bfedd", "sample_id": "a5bef5c1374b9210e4eb600b", "file": "20_seconds/HOI4D/a5bef5c1374b9210e4eb600b_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C12/N39/S346/s03/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 84/300 output frames; may include genuine out-of-view frames", "id": "006863bb291ad3c5d17bfedd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/a5bef5c1374b9210e4eb600b_MANO_GT_20.00s.mp4", "bytes": 3727987, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d1541fae0e2c8a1149f5b1c4367a2636863018b9b5a47e9842f665324d3e76ff", "mano_sha256": "a06d2c0107acd99f9ee86f5ecb17814b38229e2da0394f968fe00a81240efbde", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "04f6ecece5890272a8a32180", "sample_id": "55894168025a36d697b8ecff", "file": "20_seconds/HOI4D/55894168025a36d697b8ecff_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C2/N29/S217/s02/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 69/300 output frames; may include genuine out-of-view frames", "id": "04f6ecece5890272a8a32180", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/55894168025a36d697b8ecff_MANO_GT_20.00s.mp4", "bytes": 3464824, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "abb05f363e1eb01e3ac507aa9ebcfc74802b619c7485313766f1ee7f7e485bda", "mano_sha256": "f68a43e29366aa1ff79ec5c95511ea3ce333e5409acf530f4e5169613fcb6c10", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "0eee6840254f882b0ed6c603", "sample_id": "62234d26ca04d7fb36d69790", "file": "20_seconds/HOI4D/62234d26ca04d7fb36d69790_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C5/N46/S57/s05/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "0eee6840254f882b0ed6c603", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/62234d26ca04d7fb36d69790_MANO_GT_20.00s.mp4", "bytes": 5555159, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "144ccf4deee1c3223e5923a236a22a63042d937af43d72fe5ac68cc7698b748f", "mano_sha256": "bf9f619f87728df9073acbc44a5408a6fb2b61d3a92bd8332a851686bb8fc645", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "127085681a94c1e8a80c08bd", "sample_id": "946e599ebfd104ca502d0ef7", "file": "20_seconds/HOI4D/946e599ebfd104ca502d0ef7_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C5/N16/S263/s02/T3", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 55/300 output frames; may include genuine out-of-view frames", "id": "127085681a94c1e8a80c08bd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/946e599ebfd104ca502d0ef7_MANO_GT_20.00s.mp4", "bytes": 3287311, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "557be2523d456bb7fa70ff0d03f218542c24254aec29aff7f56aa732cf548b6e", "mano_sha256": "62a2063e41a4e9f2eb181fe5cff663a96f3415ffa5ca04b9e45dd4e87d4d509e", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "146f30c048fefadfbb647e42", "sample_id": "6dc1c4d5c2e15ae8d0ddb2fc", "file": "20_seconds/HOI4D/6dc1c4d5c2e15ae8d0ddb2fc_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C3/N46/S284/s02/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 63/300 output frames; may include genuine out-of-view frames", "id": "146f30c048fefadfbb647e42", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/6dc1c4d5c2e15ae8d0ddb2fc_MANO_GT_20.00s.mp4", "bytes": 3549336, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "354935516cc5fcac4740aff7b6156068c026c644ebd1964e5c481b94216ec3eb", "mano_sha256": "ae5db679f02ff57c3bedd6a5cfb660b664e2fbc5cf25149fc616d624dd9a300c", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "234fba7f007f269ed03f5e8e", "sample_id": "547d453246607f0deeba5058", "file": "20_seconds/HOI4D/547d453246607f0deeba5058_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C13/N11/S124/s03/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 65/300 output frames; may include genuine out-of-view frames", "id": "234fba7f007f269ed03f5e8e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/547d453246607f0deeba5058_MANO_GT_20.00s.mp4", "bytes": 3554801, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c6ed39015bb777c8d4fd6a88771a7ef0ab03b449f8a0b2c64e108fab51a8e7b2", "mano_sha256": "112b6c057a0cf1be6bbee42f0f15be524d1636ada8946189bf77c2f512c8d85b", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "2d199698da57dd199b3e1889", "sample_id": "d9f7f746570d37ceee8a9ccc", "file": "20_seconds/HOI4D/d9f7f746570d37ceee8a9ccc_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C2/N13/S212/s04/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 67/300 output frames; may include genuine out-of-view frames", "id": "2d199698da57dd199b3e1889", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/d9f7f746570d37ceee8a9ccc_MANO_GT_20.00s.mp4", "bytes": 3206400, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ef65cfc3564c556f71134c609a9ec2247f1825225579280c6bd2358de7171b11", "mano_sha256": "c230cbf444c2e38316a65150728bd39d0fe398ecc7dff45b9199c6904e1f604c", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "31b3ba85f0fbcd92930d955c", "sample_id": "5fa0075f2a0b83b010ccc9c8", "file": "20_seconds/HOI4D/5fa0075f2a0b83b010ccc9c8_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C1/N23/S104/s05/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "31b3ba85f0fbcd92930d955c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/5fa0075f2a0b83b010ccc9c8_MANO_GT_20.00s.mp4", "bytes": 3350758, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6607eba6335d4dd6d89c9cb69e4df25530417a6b628029781e25d8af66812651", "mano_sha256": "89cc93a9b2124945838eed3ded83a9762287c4fde767638c8f6d15aaced838ff", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "31f8bf63d3c17309c5329a02", "sample_id": "22a86dbd5ae802393a9162a3", "file": "20_seconds/HOI4D/22a86dbd5ae802393a9162a3_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C2/N20/S215/s05/T6", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 48/300 output frames; may include genuine out-of-view frames", "id": "31f8bf63d3c17309c5329a02", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/22a86dbd5ae802393a9162a3_MANO_GT_20.00s.mp4", "bytes": 3681366, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "97fad82bd76971998c93c4ea23fe6ca0bd18486dd849317c8a209aaec9886979", "mano_sha256": "4c6c3202dc64ab433a57423f85231fbfd7ea50f5d98070d7243dc68440c69753", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "3a6137f560b2047872d2161d", "sample_id": "7e0c24f4816ae92d46d55daa", "file": "20_seconds/HOI4D/7e0c24f4816ae92d46d55daa_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C7/N15/S270/s04/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 72/300 output frames; may include genuine out-of-view frames", "id": "3a6137f560b2047872d2161d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/7e0c24f4816ae92d46d55daa_MANO_GT_20.00s.mp4", "bytes": 4354764, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a7a9ba59f3dd540e0f118d5cf5700fc86c5ab7105d960bcac10b35e6aa0579a5", "mano_sha256": "d9a822631d8324c23adcf74edd8b36639414c3057cd0e6c36ddd65ed26069c71", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "40986b4cd3ddf09cdda0ae59", "sample_id": "1df6a39096cd7147dc406803", "file": "20_seconds/HOI4D/1df6a39096cd7147dc406803_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C14/N24/S359/s03/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 57/300 output frames; may include genuine out-of-view frames", "id": "40986b4cd3ddf09cdda0ae59", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/1df6a39096cd7147dc406803_MANO_GT_20.00s.mp4", "bytes": 3381685, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8932812cef89cbc50070feec9a25aba308b52b7e4ad8caa20a13a22218fe6c5a", "mano_sha256": "004a8f17701c9810b6666b50517468475ea4ba8a97f6c6c51b3bf38ff0bd0c14", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "519a17627bbf36991944064e", "sample_id": "64e9c6163161b953c63cc8d5", "file": "20_seconds/HOI4D/64e9c6163161b953c63cc8d5_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C20/N15/S282/s03/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 99/300 output frames; may include genuine out-of-view frames", "id": "519a17627bbf36991944064e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/64e9c6163161b953c63cc8d5_MANO_GT_20.00s.mp4", "bytes": 4366241, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8398a64d5667790219842308f5d4f06142e81a8be837e13f4384b8799e7f2077", "mano_sha256": "96c7bfe4b19b3cac7dab5f3189266005f9a4aa22e84cdfd46e8f0a74c1dcf076", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "官方手在画面边缘或画外存在位置突变告警；抽帧未见清晰手部，需核对源标签有效性，不能归因于 P3R2。", "previous_visual_review_hash_matched": "False", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "557829a9acf13ccb649711a0", "sample_id": "92e398ab9ace74ed4eacb0c7", "file": "20_seconds/HOI4D/92e398ab9ace74ed4eacb0c7_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C14/N15/S209/s02/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "hand_jump_over_25cm_per_output_frame; no valid MANO_GT mesh in 50/300 output frames; may include genuine out-of-view frames", "id": "557829a9acf13ccb649711a0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/92e398ab9ace74ed4eacb0c7_MANO_GT_20.00s.mp4", "bytes": 3168885, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b73122dc46060751265b566625ad35737a507af955b720c960a5da834ec70b4d", "mano_sha256": "89d3f38c9e18c2d1ead30fafee1354ee32cd29f36c1df90a19ee333bfe902495", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "56fa170c0234a7a2da1a32ea", "sample_id": "df8a606ae00402c499271653", "file": "20_seconds/HOI4D/df8a606ae00402c499271653_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C7/N21/S276/s03/T3", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 81/300 output frames; may include genuine out-of-view frames", "id": "56fa170c0234a7a2da1a32ea", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/df8a606ae00402c499271653_MANO_GT_20.00s.mp4", "bytes": 3801261, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d4c03a73c841366093ba3f85e6022f9ba5ae70e74b85d9575dfb7a2045f36297", "mano_sha256": "0fd9b7e2b938326e487981d1f0542a850f25388ddcbcf122869d6c93e0ac082f", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "手在部分抽帧未进入画面；原生 MANO 的有效性与遮挡需按源帧审核。", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "588adfe0aef666a4eb6c1ef2", "sample_id": "ec849ab44a19fbad5ad7a380", "file": "20_seconds/HOI4D/ec849ab44a19fbad5ad7a380_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C14/N23/S187/s01/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 45/300 output frames; may include genuine out-of-view frames", "id": "588adfe0aef666a4eb6c1ef2", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/ec849ab44a19fbad5ad7a380_MANO_GT_20.00s.mp4", "bytes": 3899078, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ce74b00de2238b49cd5d8ff0a9ea576a2fcae97fd1a1e003f3202f6f9507b53f", "mano_sha256": "48184b87c8b6d890acd3acea88dc85331e8ecc36613d04fd68c576b527f4f1ff", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "帧 141–143，官方 MANO 在画面下边缘进入并发生位置突变；需核对源标注有效性与遮挡。", "previous_visual_review_hash_matched": "False", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "5d3e4aa4dd01c8f23fc0cc25", "sample_id": "49f3d59c98c31b8291ecd181", "file": "20_seconds/HOI4D/49f3d59c98c31b8291ecd181_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C13/N32/S200/s02/T3", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "hand_jump_over_25cm_per_output_frame; no valid MANO_GT mesh in 67/300 output frames; may include genuine out-of-view frames", "id": "5d3e4aa4dd01c8f23fc0cc25", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/49f3d59c98c31b8291ecd181_MANO_GT_20.00s.mp4", "bytes": 3691185, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a30090460188e627164594a463b0d7ceeace7ba6cdcbacc6ad6a85402ce1f3e7", "mano_sha256": "68deafd88b9be6c90e4c55148dc1c9c987fcb69b845e2e10a0e091e234f8d872", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "5f4839fd682db8f840379cae", "sample_id": "9dc9ec89b5a81d85cb976ed7", "file": "20_seconds/HOI4D/9dc9ec89b5a81d85cb976ed7_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C13/N15/S346/s01/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 91/300 output frames; may include genuine out-of-view frames", "id": "5f4839fd682db8f840379cae", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/9dc9ec89b5a81d85cb976ed7_MANO_GT_20.00s.mp4", "bytes": 4461990, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "9a5f4e2592d649666c57dd496a109de0864f5e757cc7b94edbfff6cb49417731", "mano_sha256": "11e66b380192ceb2a4be92c86204adfc3c78e4166c2ce442247f53949fc74bfd", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "61f622c164c2c8ed8aff0b23", "sample_id": "c131d48e611e53212b2935d7", "file": "20_seconds/HOI4D/c131d48e611e53212b2935d7_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C3/N03/S275/s05/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 68/300 output frames; may include genuine out-of-view frames", "id": "61f622c164c2c8ed8aff0b23", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/c131d48e611e53212b2935d7_MANO_GT_20.00s.mp4", "bytes": 3569292, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4928359ef68d8cfa7f5f65b542feac969c8b1019215bc2fabb010569aec5861a", "mano_sha256": "4dabd9d9b0d85035e9d013dad7c57ef2afa720d0be28197f5f4c4fec67a51c37", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "678f1017df0c188c4c2cf150", "sample_id": "6325f3fb403e88f0a829d0de", "file": "20_seconds/HOI4D/6325f3fb403e88f0a829d0de_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C7/N44/S285/s01/T6", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 50/300 output frames; may include genuine out-of-view frames", "id": "678f1017df0c188c4c2cf150", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/6325f3fb403e88f0a829d0de_MANO_GT_20.00s.mp4", "bytes": 3707753, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2ac6e7f784c0202615fbb3968730b5ca3cb34052fe9532d6981edbac3a47b7d5", "mano_sha256": "470fbd2c4a5d1cebda90d6bd90eb05cb71994c096c121d61549b8b27eb03adb8", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "71093e8e3553143702afdc2e", "sample_id": "21c23b65b7a712beb02054d7", "file": "20_seconds/HOI4D/21c23b65b7a712beb02054d7_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C7/N45/S278/s04/T4", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 74/300 output frames; may include genuine out-of-view frames", "id": "71093e8e3553143702afdc2e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/21c23b65b7a712beb02054d7_MANO_GT_20.00s.mp4", "bytes": 3885653, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "42953ebaef5ed014782cc4c9c9af667767cb3d7e56d11301208e3ec6f153b7d3", "mano_sha256": "371bfbe556ca6c05990a3d302f199807634e6b18624fc4b6796ccab52e6b8072", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "777ae0ac229129cca4778b43", "sample_id": "50228993020df2c5750d4da4", "file": "20_seconds/HOI4D/50228993020df2c5750d4da4_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C12/N23/S157/s01/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 78/300 output frames; may include genuine out-of-view frames", "id": "777ae0ac229129cca4778b43", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/50228993020df2c5750d4da4_MANO_GT_20.00s.mp4", "bytes": 5926447, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8d7e79639984bcb4961877ff1afb1cc6fcfa9be5f3a165744d8d92fc3f82268c", "mano_sha256": "10b9113c5a315a39455b13bc782a87e72248bcb2717b98a0c1a8e91b610869ae", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "7d1d44db388699c1a563195a", "sample_id": "0eb0c050accc48aaa21a9dca", "file": "20_seconds/HOI4D/0eb0c050accc48aaa21a9dca_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C1/N26/S104/s05/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "7d1d44db388699c1a563195a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/0eb0c050accc48aaa21a9dca_MANO_GT_20.00s.mp4", "bytes": 3026738, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a8864fed38679b5740585e714402b2844a513997f86ecd50fec5f8aabee2017c", "mano_sha256": "b37fa979e2496c93f9325672d1b30c23e8721a564d4ede2fe6c36f8de88c4e82", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "7e695fb5bab04ea5eb17cc80", "sample_id": "8ef9df9a8fa20f74e0567584", "file": "20_seconds/HOI4D/8ef9df9a8fa20f74e0567584_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C12/N28/S157/s02/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 62/300 output frames; may include genuine out-of-view frames", "id": "7e695fb5bab04ea5eb17cc80", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/8ef9df9a8fa20f74e0567584_MANO_GT_20.00s.mp4", "bytes": 5837904, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c410f987758d35a8e35278497f83707f19f6508533983a67264bffcdfa964f9a", "mano_sha256": "c46beaf5e9a6682dcc705c97a569cc91c6d9e684215d6785ccadee09f43ffb1b", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "81529f236a21bb5333a22615", "sample_id": "2d1ecdfdc2f4bef02d885c8b", "file": "20_seconds/HOI4D/2d1ecdfdc2f4bef02d885c8b_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C2/N21/S213/s01/T3", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 68/300 output frames; may include genuine out-of-view frames", "id": "81529f236a21bb5333a22615", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/2d1ecdfdc2f4bef02d885c8b_MANO_GT_20.00s.mp4", "bytes": 3205921, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8cc4c06afd65eaa6908719942daf9d6833d6b24f5553c2fd9ac8c6fd78606367", "mano_sha256": "7860fa94dd4400d64b4b879e2f707487b1781d86ff74b1ef3520c451afba0848", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "85d53011e15dfb37ccd76527", "sample_id": "5e9ba3e6d65362bf59924d18", "file": "20_seconds/HOI4D/5e9ba3e6d65362bf59924d18_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C2/N20/S93/s04/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "85d53011e15dfb37ccd76527", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/5e9ba3e6d65362bf59924d18_MANO_GT_20.00s.mp4", "bytes": 5407882, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b38ee5e23410ac6aaeb161753da4e78fa2d7ab4b4d29ff5a6adeae29a8fd807f", "mano_sha256": "ee4059f4de408625ddff93343d1a69e006708f2ad2a67eaa84d697fb53ac0aeb", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "87a37c0f0b8088ace6fbdaab", "sample_id": "d6ab40b89e82a83c75023719", "file": "20_seconds/HOI4D/d6ab40b89e82a83c75023719_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C20/N15/S275/s02/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 152/300 output frames; may include genuine out-of-view frames", "id": "87a37c0f0b8088ace6fbdaab", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/d6ab40b89e82a83c75023719_MANO_GT_20.00s.mp4", "bytes": 3536396, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ee49296b375024227338c8e725fc7edbd76f34568880d82295dbc458cfd2098d", "mano_sha256": "18aa7c016d56243b2180a69a859d41d483d4a1392c046f31901683cd1c7174ae", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "90b1d27a7dd768cba6c074dc", "sample_id": "53d9460b1459fca586485507", "file": "20_seconds/HOI4D/53d9460b1459fca586485507_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C12/N17/S154/s05/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 68/300 output frames; may include genuine out-of-view frames", "id": "90b1d27a7dd768cba6c074dc", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/53d9460b1459fca586485507_MANO_GT_20.00s.mp4", "bytes": 5715324, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0ea726963472220b30ff96d2bb53bf1eb28da1048c989f555f7c375251977697", "mano_sha256": "752a9716b9d1f505e3c37c025d473a679638e371da5eda1170e71fa622f1247d", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "965c10414be2e43305756fa7", "sample_id": "a97f4ebb9837b409f52eb50b", "file": "20_seconds/HOI4D/a97f4ebb9837b409f52eb50b_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C3/N03/S96/s01/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "965c10414be2e43305756fa7", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/a97f4ebb9837b409f52eb50b_MANO_GT_20.00s.mp4", "bytes": 4146627, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "27de44a6722c88a568cb7d7792cecd6d20367ff6a2177f6c82aee5b0f71d78d3", "mano_sha256": "78f9a442f49d42e4c8a9a95b16f1e201f60f759d8ad5b92631ee1e1a1e3c8041", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "9781197819cb69ed57b33188", "sample_id": "d39302b3702e336fc0461f76", "file": "20_seconds/HOI4D/d39302b3702e336fc0461f76_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C7/N12/S280/s03/T5", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 68/300 output frames; may include genuine out-of-view frames", "id": "9781197819cb69ed57b33188", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/d39302b3702e336fc0461f76_MANO_GT_20.00s.mp4", "bytes": 3715656, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7a8b95beab763125b21ac0c33b769dfa31c9e6f36403ffd871e1f802a8f55217", "mano_sha256": "e3f8a78ac24f29b09eee0e35faedaeff8e8d2d5abc94c34dfbc2d620480207cd", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "99febb767a2868676cb52399", "sample_id": "7e809f6d9b57f24240d31dd6", "file": "20_seconds/HOI4D/7e809f6d9b57f24240d31dd6_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C12/N36/S346/s03/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 65/300 output frames; may include genuine out-of-view frames", "id": "99febb767a2868676cb52399", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/7e809f6d9b57f24240d31dd6_MANO_GT_20.00s.mp4", "bytes": 3637625, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0c99aa1bdc4764b2f6afba905a68ecc38e824013d39e3dfd2410de03947f771b", "mano_sha256": "3c91e4ec6ed13901cda951608ebe8b883e52f00360c92067ecf1044f5267b6f2", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "9d447b9bd49f04c3a07cd4fe", "sample_id": "0915528b0007edb53441c927", "file": "20_seconds/HOI4D/0915528b0007edb53441c927_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C7/N47/S278/s05/T4", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 79/300 output frames; may include genuine out-of-view frames", "id": "9d447b9bd49f04c3a07cd4fe", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/0915528b0007edb53441c927_MANO_GT_20.00s.mp4", "bytes": 3906784, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "bdbcc485ed395f1c76d3dae91ab4d69097a63437702668e0e6c81a9ec8e19098", "mano_sha256": "67dc0f49a4db4a58b5e591132bb7e290328822b0d019a5685b218208c87e290c", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a266f718112ab7cd2164e30e", "sample_id": "d78bcbb05fb0ac48c42f2868", "file": "20_seconds/HOI4D/d78bcbb05fb0ac48c42f2868_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C12/N13/S154/s04/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 60/300 output frames; may include genuine out-of-view frames", "id": "a266f718112ab7cd2164e30e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/d78bcbb05fb0ac48c42f2868_MANO_GT_20.00s.mp4", "bytes": 5805305, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d2120489c2d11a0241507519f8e7a8fd7a000f6253a199a491e6f7e478ea9b4d", "mano_sha256": "2536fe5b2fe6565c0bf8cfe0455badac18d12ed51dabd9dc1aaa62f97883d535", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a3a20e2ba5d40113de9bf996", "sample_id": "4a03356a3fe0efea423e09ce", "file": "20_seconds/HOI4D/4a03356a3fe0efea423e09ce_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C3/N08/S277/s01/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 65/300 output frames; may include genuine out-of-view frames", "id": "a3a20e2ba5d40113de9bf996", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/4a03356a3fe0efea423e09ce_MANO_GT_20.00s.mp4", "bytes": 3671745, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "90c5fa3f26203106bb05b3e7f49a6f9c64791ff6d5e2dfbe95a943b2abb782ab", "mano_sha256": "3c16be4cabbdbe9eb82a2120e3eac1980d290fa78ed6af1c8cd9532b6c39a554", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "帧 15–17，官方 MANO 在下边缘出现并跳变；保留作为标签审核候选，未自动判错。", "previous_visual_review_hash_matched": "False", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a8a65f0c11bd3c4872806488", "sample_id": "9d52176954c3397a081f9cd6", "file": "20_seconds/HOI4D/9d52176954c3397a081f9cd6_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C14/N36/S333/s03/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "hand_jump_over_25cm_per_output_frame; no valid MANO_GT mesh in 64/300 output frames; may include genuine out-of-view frames", "id": "a8a65f0c11bd3c4872806488", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/9d52176954c3397a081f9cd6_MANO_GT_20.00s.mp4", "bytes": 3222046, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e8a27f4d8dd9238f28c57a1e24cf0166619fbbcf0e57fb6c7ed51a16bfb85b77", "mano_sha256": "11632a45eef227395785a8d52ff4da295db1347645b60ce4c68fc0d589224f3d", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "af3f3f0029f4b93541c827ca", "sample_id": "340daf290b847455a27761c9", "file": "20_seconds/HOI4D/340daf290b847455a27761c9_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C5/N14/S255/s04/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 81/300 output frames; may include genuine out-of-view frames", "id": "af3f3f0029f4b93541c827ca", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/340daf290b847455a27761c9_MANO_GT_20.00s.mp4", "bytes": 3084573, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6bbcb3199b3c24d04156f4048505151d86b02a64345257f3a5c4c0865ef66877", "mano_sha256": "d36b2a5c230190939307a087889ccb592146cfb1bade61b73d7f31213a5bbbd7", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "b03f18fd1bbfaba335d934ce", "sample_id": "3e7be9ea1cc55d512f232a88", "file": "20_seconds/HOI4D/3e7be9ea1cc55d512f232a88_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C7/N46/S285/s02/T6", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 50/300 output frames; may include genuine out-of-view frames", "id": "b03f18fd1bbfaba335d934ce", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/3e7be9ea1cc55d512f232a88_MANO_GT_20.00s.mp4", "bytes": 3713503, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3fba3b1fe7af0ff630472ff0eacef49b05c491e4f407de2f7b74b172e3522184", "mano_sha256": "f379c888623d591865cebe9362ae8a819e84d4307db2d53704d99a4131595bd7", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "b2daa9eaf3fbaeb2216be432", "sample_id": "f5c9e2ae279f0d22fb4b84df", "file": "20_seconds/HOI4D/f5c9e2ae279f0d22fb4b84df_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C12/N31/S346/s01/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 80/300 output frames; may include genuine out-of-view frames", "id": "b2daa9eaf3fbaeb2216be432", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/f5c9e2ae279f0d22fb4b84df_MANO_GT_20.00s.mp4", "bytes": 4079691, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "16ecf8a3b083ccbde88f5c990c8ba2d87d41e076f35735d2b278f24979dad376", "mano_sha256": "6ab186ab6dd01671912672e46a49c3d4c7906a61289821b6e0b150d361654bb6", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "b9ec26625872109f917b0218", "sample_id": "a0b32011d35439d5d6c62d75", "file": "20_seconds/HOI4D/a0b32011d35439d5d6c62d75_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C5/N14/S121/s02/T3", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "b9ec26625872109f917b0218", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/a0b32011d35439d5d6c62d75_MANO_GT_20.00s.mp4", "bytes": 5255861, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5a0123e636c2b1eccc89147111705386406f0ece12ae71ece27be6f101d5b880", "mano_sha256": "3a2c99e8cac95d4d1c09fb770a7522e62ddf3acd3f3dafe5d79e08f2b109a909", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "cabad22d69e6848a3d7f5594", "sample_id": "adbf661805214900d33cda33", "file": "20_seconds/HOI4D/adbf661805214900d33cda33_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C14/N35/S166/s03/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 51/300 output frames; may include genuine out-of-view frames", "id": "cabad22d69e6848a3d7f5594", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/adbf661805214900d33cda33_MANO_GT_20.00s.mp4", "bytes": 2464538, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3470d1e09a3fdebc824f3f5ce39f41dde09e0c006eea5e4b67ab022eec93ab59", "mano_sha256": "5e22b05c640d8eaa957ed770686877592c6b832145b73e7b16808fbca819d559", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "d1d02ae0ca7da6a1596b4689", "sample_id": "2dc15cde3d1ce64812b08daf", "file": "20_seconds/HOI4D/2dc15cde3d1ce64812b08daf_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C2/N48/S54/s03/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "d1d02ae0ca7da6a1596b4689", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/2dc15cde3d1ce64812b08daf_MANO_GT_20.00s.mp4", "bytes": 5888494, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3806e09e9b11fc2ababb992fa0319c369a201c92441a21a11f4141c03a3c2629", "mano_sha256": "ab7465db357b162cef6dc993c5574ee880b9a53ffaed78f6098e78a2268757ff", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "d3dcc3427f0c2ce7ba63d921", "sample_id": "502868da9d35dc30aaae0fcb", "file": "20_seconds/HOI4D/502868da9d35dc30aaae0fcb_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C7/N24/S79/s04/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "d3dcc3427f0c2ce7ba63d921", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/502868da9d35dc30aaae0fcb_MANO_GT_20.00s.mp4", "bytes": 3654611, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d442897153ccb19b09d3d3f77d916fb231ddab375a4de107ae676d8900b9472d", "mano_sha256": "6f81baa6c660c951aa1d87d832abf3cb015704c14bc4933ad1c1f2cb8d1f681c", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "d494de983a8af57bd2b1575c", "sample_id": "1feb7c95071952ac5a880701", "file": "20_seconds/HOI4D/1feb7c95071952ac5a880701_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C14/N37/S333/s02/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 78/300 output frames; may include genuine out-of-view frames", "id": "d494de983a8af57bd2b1575c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/1feb7c95071952ac5a880701_MANO_GT_20.00s.mp4", "bytes": 3353749, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "34388bd36543f1fe54395321382eaa6d1dc75e674da52a0ee7ece09ae8fdba40", "mano_sha256": "17441c2c99aacf3569b8c33a98c1be1941ea245c0947fd46b0e69dc3e1a7348c", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "dadc38450dfd00553e48dc45", "sample_id": "26f8292d1afe21824fb71b22", "file": "20_seconds/HOI4D/26f8292d1afe21824fb71b22_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C3/N34/S282/s04/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 63/300 output frames; may include genuine out-of-view frames", "id": "dadc38450dfd00553e48dc45", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/26f8292d1afe21824fb71b22_MANO_GT_20.00s.mp4", "bytes": 3326266, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0246a9dfea754fc6f4facf86d3505ea897df4ef2c3b99cb6eed334e527355218", "mano_sha256": "d001f57bcbeaa90f1d725ded8097ac13f02cb1ebc0d5aeed437c05ccba0ef1d9", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "df5aa61baee11c314b2bb9ad", "sample_id": "1099941a47784070b8aa2957", "file": "20_seconds/HOI4D/1099941a47784070b8aa2957_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C12/N26/S157/s02/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 65/300 output frames; may include genuine out-of-view frames", "id": "df5aa61baee11c314b2bb9ad", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/1099941a47784070b8aa2957_MANO_GT_20.00s.mp4", "bytes": 5805280, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3db892c177c870a19f47bb6cea903dc0280994c0e413fc9760e2d9987487c34e", "mano_sha256": "a9870f932a9c0e4ffa13c3a4a62b8314e9dbcf8988cc790807df85154e35401a", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e466e4a8b5ba84946c0cfe39", "sample_id": "8d9471120e6ac0030408771a", "file": "20_seconds/HOI4D/8d9471120e6ac0030408771a_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C2/N18/S93/s03/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 19/300 output frames; may include genuine out-of-view frames", "id": "e466e4a8b5ba84946c0cfe39", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/8d9471120e6ac0030408771a_MANO_GT_20.00s.mp4", "bytes": 5876584, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "23fc6533a78abd18b0933bbdc4e06f5c86917fc0eecb0d7f44bdfdc155b1b1f1", "mano_sha256": "ed41d237b99ea88d9cc348f27addcf8638a98f67019f0d0b6aedabbb0c87223a", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e644a50dae758082dd4a8aa3", "sample_id": "7b2a34c4188adea8a44129d2", "file": "20_seconds/HOI4D/7b2a34c4188adea8a44129d2_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C2/N37/S63/s03/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "e644a50dae758082dd4a8aa3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/7b2a34c4188adea8a44129d2_MANO_GT_20.00s.mp4", "bytes": 3250256, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5274646a1b2d814efb17dd1e7b176e8822ad1241d006c3b730ddf0940aa4c817", "mano_sha256": "4364fc85caeb2af69989973ca0734e2e413c6a3d97ce8171ad63741f2d010cc9", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e7d6ace94cea80607ee4b1a0", "sample_id": "eecb8dc458454a2d7fbfb065", "file": "20_seconds/HOI4D/eecb8dc458454a2d7fbfb065_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800003/H3/C3/N46/S282/s01/T2", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 40/300 output frames; may include genuine out-of-view frames", "id": "e7d6ace94cea80607ee4b1a0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/eecb8dc458454a2d7fbfb065_MANO_GT_20.00s.mp4", "bytes": 3580837, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "edbc9b1c1bf75c28ef5e5ac56eb256656efcc26cc0d4d9367d257de2a4506617", "mano_sha256": "cb71756cef1ebac67a127b6baa46d35649a47e75bc8e5f3028cfed432d5b6331", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "f964be65b5020e10bc1b6b45", "sample_id": "9e5a4e9489a1df58dfd0f72a", "file": "20_seconds/HOI4D/9e5a4e9489a1df58dfd0f72a_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C13/N32/S128/s03/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "no valid MANO_GT mesh in 54/300 output frames; may include genuine out-of-view frames", "id": "f964be65b5020e10bc1b6b45", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/9e5a4e9489a1df58dfd0f72a_MANO_GT_20.00s.mp4", "bytes": 4690576, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "fcafd393a3ddebe068541290967afa6a597321f4c62742ad886fe966864750f9", "mano_sha256": "25c1cfaebc5ca940d989b24d81b19285fec6313a00abba1130be032b076fb35e", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "帧 288–290，官方 MANO 在画面下边缘骤然放大/截断；列为源标注异常候选，待回查原标注。", "previous_visual_review_hash_matched": "False", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "fd49b232b6ff873fbfa2f141", "sample_id": "fe633b55b4899948643adec3", "file": "20_seconds/HOI4D/fe633b55b4899948643adec3_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800002/H2/C13/N15/S124/s04/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "hand_jump_over_25cm_per_output_frame; no valid MANO_GT mesh in 65/300 output frames; may include genuine out-of-view frames", "id": "fd49b232b6ff873fbfa2f141", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/fe633b55b4899948643adec3_MANO_GT_20.00s.mp4", "bytes": 4123469, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "HOI4D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "bb7754d5797cfc564ebc22a99b290f582f6046b2d9510af724b3745701a19558", "mano_sha256": "e8bdd89a65e52936351ce8cdb428e0773d7ffc397847965fda04f7f11c2750ba", "duration_s": 20.0, "frames": 300, "fps": 15.0, "mano_source": "native HOI4D poseCoeff/beta/trans; unmodified source MANO-left convention", "scene_depth_source": "native HOI4D registered sensor depth", "scene_camera_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "source calibration", "scene_scale_anchor": "native sensor metres", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "ffbe5a3fa4fda77892e020ca", "sample_id": "4505c34060b6c250fe917d26", "file": "20_seconds/HOI4D/4505c34060b6c250fe917d26_MANO_GT_20.00s.mp4", "recording_id": "ZY20210800001/H1/C7/N28/S60/s02/T1", "source_first_frame": "0", "source_last_frame": "299", "source_split": "external_train", "training_label_source": "missing", "quality_warnings": "", "id": "ffbe5a3fa4fda77892e020ca", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HOI4D/4505c34060b6c250fe917d26_MANO_GT_20.00s.mp4", "bytes": 4524318, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考适配器没有有效世界系轨迹；视频中的 ViPE 相机是估计结果。" }, "camera_display_source": "historical ViPE using native sensor depth; 7.500 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "241e4b7644dc6434c18744865f1dc21bc198f7f5f06dfc1628bc480ed36ba112", "mano_sha256": "149efd2cb17ea0b811a96d86a15fbb4c5c9aad19a5b146dd031e20073b49d51b", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "00b2632b66fcb00939ecb327", "sample_id": "3517be8c1b478bf8628efa27", "file": "20_seconds/ARCTIC/3517be8c1b478bf8628efa27_MANO_GT_20.00s.mp4", "recording_id": "s07/ketchup_grab_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 35/600 output frames; may include genuine out-of-view frames", "id": "00b2632b66fcb00939ecb327", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/3517be8c1b478bf8628efa27_MANO_GT_20.00s.mp4", "bytes": 4818502, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05037183667701099, "p90": 0.17678015505751185, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 8.944173938286479, "p90": 42.09647775832859, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.1935924537875911, "p90": 0.44617808316646096, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 0.6076794657762938 }, "cmh_m_s": { "median": 0.07384359369284726, "p90": 0.2670046179194657, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 0.6076794657762938 }, "rhcm_continuous_estimate_m_s": { "median": 0.16289034835134014, "p90": 0.33424393241375205, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 0.6076794657762938 }, "rhcm_discrete_m_s": { "median": 0.1628510159674013, "p90": 0.3341857539370892, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 0.6076794657762938 }, "wms": { "median": 0.6220060064693126, "p90": 0.9289975156431662, "valid_pairs": 361, "valid_duration_s": 12.033333333333335, "coverage": 0.602671118530885 }, "rham_deg_s": { "median": 54.21177436548176, "p90": 162.82687474877838, "valid_pairs": 364, "valid_duration_s": 12.133333333333333, "coverage": 0.6076794657762938 }, "speed_spearman": 0.45220433364891643, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.20830397356262287, "p90": 0.4762781688349668, "valid_pairs": 396, "valid_duration_s": 13.2, "coverage": 0.66110183639399 }, "cmh_m_s": { "median": 0.04745850726086623, "p90": 0.23047492287791432, "valid_pairs": 396, "valid_duration_s": 13.2, "coverage": 0.66110183639399 }, "rhcm_continuous_estimate_m_s": { "median": 0.17221790059142467, "p90": 0.3748154307080126, "valid_pairs": 396, "valid_duration_s": 13.2, "coverage": 0.66110183639399 }, "rhcm_discrete_m_s": { "median": 0.17215354167133623, "p90": 0.37484239462782554, "valid_pairs": 396, "valid_duration_s": 13.2, "coverage": 0.66110183639399 }, "wms": { "median": 0.7492087466099868, "p90": 0.9503418586660956, "valid_pairs": 396, "valid_duration_s": 13.2, "coverage": 0.66110183639399 }, "rham_deg_s": { "median": 45.83897025331643, "p90": 135.37189106687268, "valid_pairs": 396, "valid_duration_s": 13.2, "coverage": 0.66110183639399 }, "speed_spearman": 0.4682518196021369, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.17678015505751185, "camera_angular_speed_p90_deg_s": 42.09647775832859, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4762781688349668, "hand_camera_speed_p90_m_s": 0.37484239462782554, "hand_world_valid_fraction": 0.6391666666666667, "hand_camera_valid_fraction": 0.6391666666666667, "cmh_p90_m_s": 0.2670046179194657, "rhcm_p90_m_s": 0.37484239462782554, "rham_p90_deg_s": 162.82687474877838, "wms_p90": 0.9503418586660956, "gt_coverage": 0.66110183639399, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "17734bdcaf4ef136328f19d2b294629b162a3c2188ed21d583744e35432f3926", "metadata_sha256": "d83c8eb53931ebc4c6a9acca87438c01f507d1f4db7b11e7623614280c513657", "source_file": "long20_20260918/local/train/previews/00b2632b66fcb00939ecb327/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.660953740672398e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "404d9e4f1e9b2e409fc6568c13ab907b4a0c88f6afecb0954c2059298b890eb6", "mano_sha256": "23dc1ed9be8e6a07611a439dd4511c095a482e774b3fd891e08ea07cbdcff413", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "017706604e06ebe4ac1f2bfe", "sample_id": "3ba1b3faac68e00ce6780de2", "file": "20_seconds/ARCTIC/3ba1b3faac68e00ce6780de2_MANO_GT_20.00s.mp4", "recording_id": "s07/phone_use_01", "source_first_frame": "2", "source_last_frame": "601", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 14/600 output frames; may include genuine out-of-view frames", "id": "017706604e06ebe4ac1f2bfe", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/3ba1b3faac68e00ce6780de2_MANO_GT_20.00s.mp4", "bytes": 5166218, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0587465984806104, "p90": 0.16023376080671867, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 11.714931899395529, "p90": 30.8133226377779, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.10260220612941767, "p90": 0.2707373136154658, "valid_pairs": 368, "valid_duration_s": 12.266666666666666, "coverage": 0.6143572621035058 }, "cmh_m_s": { "median": 0.06945225713211972, "p90": 0.14156036764366078, "valid_pairs": 368, "valid_duration_s": 12.266666666666666, "coverage": 0.6143572621035058 }, "rhcm_continuous_estimate_m_s": { "median": 0.07073567681465792, "p90": 0.23262362196715092, "valid_pairs": 368, "valid_duration_s": 12.266666666666666, "coverage": 0.6143572621035058 }, "rhcm_discrete_m_s": { "median": 0.0707713033484473, "p90": 0.23263078062747589, "valid_pairs": 368, "valid_duration_s": 12.266666666666666, "coverage": 0.6143572621035058 }, "wms": { "median": 0.5037862064916256, "p90": 0.8498396668799478, "valid_pairs": 368, "valid_duration_s": 12.266666666666666, "coverage": 0.6143572621035058 }, "rham_deg_s": { "median": 50.17891966763697, "p90": 109.63893873010886, "valid_pairs": 368, "valid_duration_s": 12.266666666666666, "coverage": 0.6143572621035058 }, "speed_spearman": 0.5509423131193758, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09471334122746801, "p90": 0.24837931408904498, "valid_pairs": 445, "valid_duration_s": 14.833333333333334, "coverage": 0.7429048414023374 }, "cmh_m_s": { "median": 0.051454622432131746, "p90": 0.12905138509374625, "valid_pairs": 445, "valid_duration_s": 14.833333333333334, "coverage": 0.7429048414023374 }, "rhcm_continuous_estimate_m_s": { "median": 0.08911322179415415, "p90": 0.2173069767592392, "valid_pairs": 445, "valid_duration_s": 14.833333333333334, "coverage": 0.7429048414023374 }, "rhcm_discrete_m_s": { "median": 0.08951360631347098, "p90": 0.21734012511868453, "valid_pairs": 445, "valid_duration_s": 14.833333333333334, "coverage": 0.7429048414023374 }, "wms": { "median": 0.6406561703038363, "p90": 0.9211445666996709, "valid_pairs": 445, "valid_duration_s": 14.833333333333334, "coverage": 0.7429048414023374 }, "rham_deg_s": { "median": 69.9714152362228, "p90": 147.54714281694268, "valid_pairs": 445, "valid_duration_s": 14.833333333333334, "coverage": 0.7429048414023374 }, "speed_spearman": 0.25736002037206246, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.16023376080671867, "camera_angular_speed_p90_deg_s": 30.8133226377779, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2707373136154658, "hand_camera_speed_p90_m_s": 0.23263078062747589, "hand_world_valid_fraction": 0.6858333333333333, "hand_camera_valid_fraction": 0.6858333333333333, "cmh_p90_m_s": 0.14156036764366078, "rhcm_p90_m_s": 0.23263078062747589, "rham_p90_deg_s": 147.54714281694268, "wms_p90": 0.9211445666996709, "gt_coverage": 0.7429048414023374, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "bc60be070f65c8cd6ec870d8b871e52d67172339afb4fc51631fe805242a1cc9", "metadata_sha256": "16994d4ac95fea8c9335b61ab59528f84f8129fa5fe72dbf0366d315742ba684", "source_file": "long20_20260918/local/train/previews/017706604e06ebe4ac1f2bfe/reference.npz", "window_s": [ 0.03333333333333333, 20.0 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.577566798522237e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "19fc5b663b11fbfcf5d17bbaf0ef99edf9a5de7c2c4e0107360ba2c0e44ca4b9", "mano_sha256": "6c0591b86f8b6159b7d23703e42b5bf4e271191f71c45abb7bd5c1fc7cb2edeb", "duration_s": 19.9, "frames": 597, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "035aae7bb26853f736247f21", "sample_id": "e927a1d15409e3916d801c70", "file": "short_originals_under_20s/ARCTIC/e927a1d15409e3916d801c70_MANO_GT_19.90s.mp4", "recording_id": "s04/mixer_use_02", "source_first_frame": "2", "source_last_frame": "598", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 87/597 output frames; may include genuine out-of-view frames", "id": "035aae7bb26853f736247f21", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/ARCTIC/e927a1d15409e3916d801c70_MANO_GT_19.90s.mp4", "bytes": 5102600, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.050613038730277296, "p90": 0.08306749219634113, "valid_pairs": 596, "valid_duration_s": 19.866666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 11.009399971432565, "p90": 17.798757060172548, "valid_pairs": 596, "valid_duration_s": 19.866666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.09096654650893682, "p90": 0.28870504538594893, "valid_pairs": 509, "valid_duration_s": 16.966666666666665, "coverage": 0.8540268456375838 }, "cmh_m_s": { "median": 0.0664430745072354, "p90": 0.13327518176908523, "valid_pairs": 509, "valid_duration_s": 16.966666666666665, "coverage": 0.8540268456375838 }, "rhcm_continuous_estimate_m_s": { "median": 0.1177487986087658, "p90": 0.36101336702793896, "valid_pairs": 509, "valid_duration_s": 16.966666666666665, "coverage": 0.8540268456375838 }, "rhcm_discrete_m_s": { "median": 0.11769714012802614, "p90": 0.3611957047776099, "valid_pairs": 509, "valid_duration_s": 16.966666666666665, "coverage": 0.8540268456375838 }, "wms": { "median": 0.7579100087114115, "p90": 0.9595386556546858, "valid_pairs": 509, "valid_duration_s": 16.966666666666665, "coverage": 0.8540268456375838 }, "rham_deg_s": { "median": 48.344577011826786, "p90": 145.47194904150746, "valid_pairs": 509, "valid_duration_s": 16.966666666666665, "coverage": 0.8540268456375838 }, "speed_spearman": 0.4679298336292271, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1965434421500597, "p90": 0.44144717642251285, "valid_pairs": 486, "valid_duration_s": 16.2, "coverage": 0.8154362416107382 }, "cmh_m_s": { "median": 0.06362552765992083, "p90": 0.11436224544948453, "valid_pairs": 486, "valid_duration_s": 16.2, "coverage": 0.8154362416107382 }, "rhcm_continuous_estimate_m_s": { "median": 0.2187576939839307, "p90": 0.4841386015415946, "valid_pairs": 486, "valid_duration_s": 16.2, "coverage": 0.8154362416107382 }, "rhcm_discrete_m_s": { "median": 0.21877060619384192, "p90": 0.4839518746162531, "valid_pairs": 486, "valid_duration_s": 16.2, "coverage": 0.8154362416107382 }, "wms": { "median": 0.8671135479134862, "p90": 0.9766872746697568, "valid_pairs": 486, "valid_duration_s": 16.2, "coverage": 0.8154362416107382 }, "rham_deg_s": { "median": 66.04664856308307, "p90": 153.64021651865698, "valid_pairs": 486, "valid_duration_s": 16.2, "coverage": 0.8154362416107382 }, "speed_spearman": 0.19493680545918451, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.08306749219634113, "camera_angular_speed_p90_deg_s": 17.798757060172548, "camera_valid_fraction": 1.0, "camera_pairs": 596, "hand_world_speed_p90_m_s": 0.44144717642251285, "hand_camera_speed_p90_m_s": 0.4839518746162531, "hand_world_valid_fraction": 0.8366834170854272, "hand_camera_valid_fraction": 0.8366834170854272, "cmh_p90_m_s": 0.13327518176908523, "rhcm_p90_m_s": 0.4839518746162531, "rham_p90_deg_s": 153.64021651865698, "wms_p90": 0.9766872746697568, "gt_coverage": 0.8540268456375838, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "7bb27712a536f24fca505caf0694a51ce55256d1e70b86a5430225ea260ae8eb", "metadata_sha256": "72e2b59130fa1bdd2fd76da5b5eef178bf310ebac2428ff970e7657afda28c41", "source_file": "long20_20260918/local/train/previews/035aae7bb26853f736247f21/reference.npz", "window_s": [ 0.0, 19.866666666666667 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 597, "common_world_residual_max_m": 4.847799515902576e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b694e4941ae2e1cfdbfa7d1082bcc0aaaa5a599b30e3ad3e2dfd81ccafa25f24", "mano_sha256": "bf37902574838fcd9287e428869555819839743d699e3905533b3631a8b936bf", "duration_s": 19.134, "frames": 574, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "05fca95a1079712f2201378a", "sample_id": "0896860b6b989ff9d536d262", "file": "short_originals_under_20s/ARCTIC/0896860b6b989ff9d536d262_MANO_GT_19.13s.mp4", "recording_id": "s02/box_grab_01", "source_first_frame": "2", "source_last_frame": "575", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 182/574 output frames; may include genuine out-of-view frames", "id": "05fca95a1079712f2201378a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/ARCTIC/0896860b6b989ff9d536d262_MANO_GT_19.13s.mp4", "bytes": 4006458, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05206330354328276, "p90": 0.11415047696515751, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 1.0 }, "camera_rotation": { "median": 6.315863452645461, "p90": 13.292886281818996, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.19708471267776534, "p90": 0.4513574202431941, "valid_pairs": 369, "valid_duration_s": 12.299999999999999, "coverage": 0.6439790575916229 }, "cmh_m_s": { "median": 0.055753502392255065, "p90": 0.10758282794647661, "valid_pairs": 369, "valid_duration_s": 12.299999999999999, "coverage": 0.6439790575916229 }, "rhcm_continuous_estimate_m_s": { "median": 0.1787019909455187, "p90": 0.43430367011834353, "valid_pairs": 369, "valid_duration_s": 12.299999999999999, "coverage": 0.6439790575916229 }, "rhcm_discrete_m_s": { "median": 0.17870271602630625, "p90": 0.4343332448313281, "valid_pairs": 369, "valid_duration_s": 12.299999999999999, "coverage": 0.6439790575916229 }, "wms": { "median": 0.7392649416649636, "p90": 0.9195447081380649, "valid_pairs": 369, "valid_duration_s": 12.299999999999999, "coverage": 0.6439790575916229 }, "rham_deg_s": { "median": 29.861181849711983, "p90": 67.56019824132593, "valid_pairs": 369, "valid_duration_s": 12.299999999999999, "coverage": 0.6439790575916229 }, "speed_spearman": 0.30028660686136827, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.23675908011682645, "p90": 0.5574541323997879, "valid_pairs": 362, "valid_duration_s": 12.066666666666666, "coverage": 0.631762652705061 }, "cmh_m_s": { "median": 0.07123385335510371, "p90": 0.16038545559744352, "valid_pairs": 362, "valid_duration_s": 12.066666666666666, "coverage": 0.631762652705061 }, "rhcm_continuous_estimate_m_s": { "median": 0.19798578025087654, "p90": 0.509206649454282, "valid_pairs": 362, "valid_duration_s": 12.066666666666666, "coverage": 0.631762652705061 }, "rhcm_discrete_m_s": { "median": 0.1980737212198388, "p90": 0.5092130179754184, "valid_pairs": 362, "valid_duration_s": 12.066666666666666, "coverage": 0.631762652705061 }, "wms": { "median": 0.691084570733348, "p90": 0.8842027858822815, "valid_pairs": 362, "valid_duration_s": 12.066666666666666, "coverage": 0.631762652705061 }, "rham_deg_s": { "median": 34.144610568544465, "p90": 100.66160932296953, "valid_pairs": 362, "valid_duration_s": 12.066666666666666, "coverage": 0.631762652705061 }, "speed_spearman": 0.31625505406411447, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.11415047696515751, "camera_angular_speed_p90_deg_s": 13.292886281818996, "camera_valid_fraction": 1.0, "camera_pairs": 573, "hand_world_speed_p90_m_s": 0.5574541323997879, "hand_camera_speed_p90_m_s": 0.5092130179754184, "hand_world_valid_fraction": 0.6437282229965157, "hand_camera_valid_fraction": 0.6437282229965157, "cmh_p90_m_s": 0.16038545559744352, "rhcm_p90_m_s": 0.5092130179754184, "rham_p90_deg_s": 100.66160932296953, "wms_p90": 0.9195447081380649, "gt_coverage": 0.6439790575916229, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "980f5e2df42479e84d376c90e1f8ccc753a207dc9627c5bf14ea33138e229580", "metadata_sha256": "d1ef4a50a35c98803a1641b90a8294af4898eb1509d5d5cecb5849f53ab6dcd1", "source_file": "long20_20260918/local/train/previews/05fca95a1079712f2201378a/reference.npz", "window_s": [ 0.0, 19.1 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 574, "common_world_residual_max_m": 4.518280359883027e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6c4bf6b52d8f9a1176314492e3d41784f6d386de9eeb78de513963559ab2fd51", "mano_sha256": "230d45869ada73d278926cf06122e3a35948b79afd1586316de4ed907b553d95", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "0633f807290f2888d081efcc", "sample_id": "20fb61bb012f3438adf409e2", "file": "20_seconds/ARCTIC/20fb61bb012f3438adf409e2_MANO_GT_20.00s.mp4", "recording_id": "s06/espressomachine_grab_01", "source_first_frame": "166", "source_last_frame": "765", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 40/600 output frames; may include genuine out-of-view frames", "id": "0633f807290f2888d081efcc", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/20fb61bb012f3438adf409e2_MANO_GT_20.00s.mp4", "bytes": 4714072, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05770988752166027, "p90": 0.1321582807775684, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 8.489319053536295, "p90": 15.401809498645969, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.19227623518958376, "p90": 0.6674430634940205, "valid_pairs": 351, "valid_duration_s": 11.7, "coverage": 0.5859766277128547 }, "cmh_m_s": { "median": 0.07802109370725271, "p90": 0.14321843643047785, "valid_pairs": 351, "valid_duration_s": 11.7, "coverage": 0.5859766277128547 }, "rhcm_continuous_estimate_m_s": { "median": 0.21357119313862086, "p90": 0.7389395381767664, "valid_pairs": 351, "valid_duration_s": 11.7, "coverage": 0.5859766277128547 }, "rhcm_discrete_m_s": { "median": 0.21360930572264442, "p90": 0.7391937398985476, "valid_pairs": 351, "valid_duration_s": 11.7, "coverage": 0.5859766277128547 }, "wms": { "median": 0.7809108204026391, "p90": 0.9565686836413518, "valid_pairs": 351, "valid_duration_s": 11.7, "coverage": 0.5859766277128547 }, "rham_deg_s": { "median": 41.69552411566647, "p90": 165.04572302037144, "valid_pairs": 351, "valid_duration_s": 11.7, "coverage": 0.5859766277128547 }, "speed_spearman": 0.1442948717948718, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.15668053355020223, "p90": 0.4500443934245079, "valid_pairs": 463, "valid_duration_s": 15.433333333333335, "coverage": 0.7729549248747914 }, "cmh_m_s": { "median": 0.07518162228887361, "p90": 0.1371237738129986, "valid_pairs": 463, "valid_duration_s": 15.433333333333335, "coverage": 0.7729549248747914 }, "rhcm_continuous_estimate_m_s": { "median": 0.15251725865131704, "p90": 0.40326664296716996, "valid_pairs": 463, "valid_duration_s": 15.433333333333335, "coverage": 0.7729549248747914 }, "rhcm_discrete_m_s": { "median": 0.15250092575013968, "p90": 0.40301318128624575, "valid_pairs": 463, "valid_duration_s": 15.433333333333335, "coverage": 0.7729549248747914 }, "wms": { "median": 0.6846702059875179, "p90": 0.9183578759023043, "valid_pairs": 463, "valid_duration_s": 15.433333333333335, "coverage": 0.7729549248747914 }, "rham_deg_s": { "median": 37.670528632166814, "p90": 131.50212374597965, "valid_pairs": 463, "valid_duration_s": 15.433333333333335, "coverage": 0.7729549248747914 }, "speed_spearman": 0.1916583081772625, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1321582807775684, "camera_angular_speed_p90_deg_s": 15.401809498645969, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6674430634940205, "hand_camera_speed_p90_m_s": 0.7391937398985476, "hand_world_valid_fraction": 0.6833333333333333, "hand_camera_valid_fraction": 0.6833333333333333, "cmh_p90_m_s": 0.14321843643047785, "rhcm_p90_m_s": 0.7391937398985476, "rham_p90_deg_s": 165.04572302037144, "wms_p90": 0.9565686836413518, "gt_coverage": 0.7729549248747914, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "be9600bb931f3e75d0944d8da438a4bce18b6b44688f2a3756f9fe99ec9c682e", "metadata_sha256": "a5938b0a07c13764dd69956052ab014792c1b91962219bfa6cc76441019ae47e", "source_file": "long20_20260918/local/train/previews/0633f807290f2888d081efcc/reference.npz", "window_s": [ 5.5, 25.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.660953740672398e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ad077c76bcab4ac5824bdffe79065cf24a45779bce4d5658ca4e6e759fbffecb", "mano_sha256": "fdb458b7f37106068eafcc203e5eb9ca09b56ec2a4fac9f044a369d4250027bf", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "0e8b6eff88b7698cfe619f3b", "sample_id": "a427e952c25f8f03fb543d52", "file": "20_seconds/ARCTIC/a427e952c25f8f03fb543d52_MANO_GT_20.00s.mp4", "recording_id": "s05/ketchup_use_01", "source_first_frame": "178", "source_last_frame": "777", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 55/600 output frames; may include genuine out-of-view frames", "id": "0e8b6eff88b7698cfe619f3b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/a427e952c25f8f03fb543d52_MANO_GT_20.00s.mp4", "bytes": 5927311, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08455410989874347, "p90": 0.1979691312590607, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 17.647594903902885, "p90": 48.58699241338098, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.14385427320475366, "p90": 0.3539552420145938, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193656 }, "cmh_m_s": { "median": 0.10282632718722548, "p90": 0.2213204491219491, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193656 }, "rhcm_continuous_estimate_m_s": { "median": 0.15999421946462494, "p90": 0.3810768996728613, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193656 }, "rhcm_discrete_m_s": { "median": 0.15985304446306906, "p90": 0.38195880850747566, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193656 }, "wms": { "median": 0.7125419646040586, "p90": 0.9532210909422391, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193656 }, "rham_deg_s": { "median": 82.6601079323058, "p90": 198.7447462438405, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193656 }, "speed_spearman": 0.3264698744677565, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.12103183921758354, "p90": 0.3491651230417097, "valid_pairs": 520, "valid_duration_s": 17.333333333333336, "coverage": 0.8681135225375627 }, "cmh_m_s": { "median": 0.09115129278060251, "p90": 0.21744917143710946, "valid_pairs": 520, "valid_duration_s": 17.333333333333336, "coverage": 0.8681135225375627 }, "rhcm_continuous_estimate_m_s": { "median": 0.14654162214301775, "p90": 0.45992416539668807, "valid_pairs": 520, "valid_duration_s": 17.333333333333336, "coverage": 0.8681135225375627 }, "rhcm_discrete_m_s": { "median": 0.14733393158312708, "p90": 0.4579392817469557, "valid_pairs": 520, "valid_duration_s": 17.333333333333336, "coverage": 0.8681135225375627 }, "wms": { "median": 0.7570041490103956, "p90": 0.9449385262735001, "valid_pairs": 520, "valid_duration_s": 17.333333333333336, "coverage": 0.8681135225375627 }, "rham_deg_s": { "median": 79.42688953349642, "p90": 244.607060758342, "valid_pairs": 520, "valid_duration_s": 17.333333333333336, "coverage": 0.8681135225375627 }, "speed_spearman": 0.4927796444399685, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1979691312590607, "camera_angular_speed_p90_deg_s": 48.58699241338098, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3539552420145938, "hand_camera_speed_p90_m_s": 0.4579392817469557, "hand_world_valid_fraction": 0.87, "hand_camera_valid_fraction": 0.87, "cmh_p90_m_s": 0.2213204491219491, "rhcm_p90_m_s": 0.4579392817469557, "rham_p90_deg_s": 244.607060758342, "wms_p90": 0.9532210909422391, "gt_coverage": 0.8697829716193656, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "a63fc892eec8061a98f605cc16f71c4cc29d7c53dae429957d34c5e6295fd7b7", "metadata_sha256": "4e8811658a1dcda07ff2c008e223c6f106d74f7f0a6f9e0db4ed7ed9fcff73b8", "source_file": "long20_20260918/local/benchmark/previews/0e8b6eff88b7698cfe619f3b/reference.npz", "window_s": [ 5.9, 25.866666666666667 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.75098925995612e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "dbafe6f28ed838280a94e3ec2f663217489aa6f9ce93d72187c50b0ffa4e69b0", "mano_sha256": "3f35ee67d7020e025b5c8c271d4d8a2214bd9cbec6bed8f06ed404e647a32eb8", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "1030f7521b9f9a20d08a3dc9", "sample_id": "bbae20ebcb66954587372a88", "file": "20_seconds/ARCTIC/bbae20ebcb66954587372a88_MANO_GT_20.00s.mp4", "recording_id": "s02/notebook_use_02", "source_first_frame": "107", "source_last_frame": "706", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 57/600 output frames; may include genuine out-of-view frames", "id": "1030f7521b9f9a20d08a3dc9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/bbae20ebcb66954587372a88_MANO_GT_20.00s.mp4", "bytes": 5366795, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.021783724369828606, "p90": 0.05267610773994911, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 4.39113160411801, "p90": 8.941678467852327, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03647554898778423, "p90": 0.2557419378475425, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554259 }, "cmh_m_s": { "median": 0.028818300256435926, "p90": 0.0544089598757259, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554259 }, "rhcm_continuous_estimate_m_s": { "median": 0.04351525918767129, "p90": 0.2420766657036255, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554259 }, "rhcm_discrete_m_s": { "median": 0.04351642955699771, "p90": 0.24208611330858612, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554259 }, "wms": { "median": 0.6887224598855397, "p90": 0.9224581204399619, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.898163606010017 }, "rham_deg_s": { "median": 15.29638496321523, "p90": 97.87835455767463, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554259 }, "speed_spearman": 0.20005676396874872, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.20641951128039251, "p90": 0.429498626191223, "valid_pairs": 530, "valid_duration_s": 17.666666666666664, "coverage": 0.8848080133555926 }, "cmh_m_s": { "median": 0.025486302219082192, "p90": 0.0474882773769507, "valid_pairs": 530, "valid_duration_s": 17.666666666666664, "coverage": 0.8848080133555926 }, "rhcm_continuous_estimate_m_s": { "median": 0.20426852936674927, "p90": 0.4383014297095848, "valid_pairs": 530, "valid_duration_s": 17.666666666666664, "coverage": 0.8848080133555926 }, "rhcm_discrete_m_s": { "median": 0.20432682904473665, "p90": 0.4383136225854418, "valid_pairs": 530, "valid_duration_s": 17.666666666666664, "coverage": 0.8848080133555926 }, "wms": { "median": 0.9056328853085839, "p90": 0.984712694142267, "valid_pairs": 530, "valid_duration_s": 17.666666666666664, "coverage": 0.8848080133555926 }, "rham_deg_s": { "median": 65.36674230839247, "p90": 136.04062700538813, "valid_pairs": 530, "valid_duration_s": 17.666666666666664, "coverage": 0.8848080133555926 }, "speed_spearman": 0.10940849148290525, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05267610773994911, "camera_angular_speed_p90_deg_s": 8.941678467852327, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.429498626191223, "hand_camera_speed_p90_m_s": 0.4383136225854418, "hand_world_valid_fraction": 0.895, "hand_camera_valid_fraction": 0.895, "cmh_p90_m_s": 0.0544089598757259, "rhcm_p90_m_s": 0.4383136225854418, "rham_p90_deg_s": 136.04062700538813, "wms_p90": 0.984712694142267, "gt_coverage": 0.9031719532554259, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "4c86a60602fafe838eb904349e06313a686135701857c88f773f77009ddfabe0", "metadata_sha256": "aa37fe25a44cea6359ec01c49cf372d122b407e39f43a35498c7835f2365dc66", "source_file": "long20_20260918/local/train/previews/1030f7521b9f9a20d08a3dc9/reference.npz", "window_s": [ 3.5, 23.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.509747244882934e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "523a77d2f831ca62838709e70e90660bdc2a8210a090276368431d33af91a751", "mano_sha256": "c63930bf676447a339a257e28b90c829d0c5997608dfe14d4466596a74347c9e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "18297a653ce83463b04bc637", "sample_id": "f3ecc1895a3f758b0192d153", "file": "20_seconds/ARCTIC/f3ecc1895a3f758b0192d153_MANO_GT_20.00s.mp4", "recording_id": "s07/waffleiron_use_02", "source_first_frame": "3", "source_last_frame": "602", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 18/600 output frames; may include genuine out-of-view frames", "id": "18297a653ce83463b04bc637", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/f3ecc1895a3f758b0192d153_MANO_GT_20.00s.mp4", "bytes": 5118203, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0460813419474033, "p90": 0.08274213745488333, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 6.5986118001346, "p90": 11.241030124537359, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.1079079407563155, "p90": 0.32371901645613466, "valid_pairs": 581, "valid_duration_s": 19.366666666666667, "coverage": 0.9699499165275458 }, "cmh_m_s": { "median": 0.043002239542826254, "p90": 0.08119427159387739, "valid_pairs": 581, "valid_duration_s": 19.366666666666667, "coverage": 0.9699499165275458 }, "rhcm_continuous_estimate_m_s": { "median": 0.10255448440373649, "p90": 0.2970266239883701, "valid_pairs": 581, "valid_duration_s": 19.366666666666667, "coverage": 0.9699499165275458 }, "rhcm_discrete_m_s": { "median": 0.10259955171220049, "p90": 0.2971243370534796, "valid_pairs": 581, "valid_duration_s": 19.366666666666667, "coverage": 0.9699499165275458 }, "wms": { "median": 0.7179961112423238, "p90": 0.9411637986403157, "valid_pairs": 581, "valid_duration_s": 19.366666666666667, "coverage": 0.9699499165275458 }, "rham_deg_s": { "median": 53.3141864869546, "p90": 122.93665507983377, "valid_pairs": 581, "valid_duration_s": 19.366666666666667, "coverage": 0.9699499165275458 }, "speed_spearman": 0.30215192189202705, "hand_class": "high", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.13878372057653424, "p90": 0.36545950262638877, "valid_pairs": 497, "valid_duration_s": 16.56666666666667, "coverage": 0.8297161936560936 }, "cmh_m_s": { "median": 0.04589048554519244, "p90": 0.08936527090651693, "valid_pairs": 497, "valid_duration_s": 16.56666666666667, "coverage": 0.8297161936560936 }, "rhcm_continuous_estimate_m_s": { "median": 0.13894994297083696, "p90": 0.3304895120518767, "valid_pairs": 497, "valid_duration_s": 16.56666666666667, "coverage": 0.8297161936560936 }, "rhcm_discrete_m_s": { "median": 0.1388666889573243, "p90": 0.3303257845006308, "valid_pairs": 497, "valid_duration_s": 16.56666666666667, "coverage": 0.8297161936560936 }, "wms": { "median": 0.7361475271729055, "p90": 0.9674313822178536, "valid_pairs": 497, "valid_duration_s": 16.56666666666667, "coverage": 0.8297161936560936 }, "rham_deg_s": { "median": 51.29595845638088, "p90": 148.8857078080165, "valid_pairs": 497, "valid_duration_s": 16.56666666666667, "coverage": 0.8297161936560936 }, "speed_spearman": 0.21048297167380497, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.08274213745488333, "camera_angular_speed_p90_deg_s": 11.241030124537359, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.36545950262638877, "hand_camera_speed_p90_m_s": 0.3303257845006308, "hand_world_valid_fraction": 0.905, "hand_camera_valid_fraction": 0.905, "cmh_p90_m_s": 0.08936527090651693, "rhcm_p90_m_s": 0.3303257845006308, "rham_p90_deg_s": 148.8857078080165, "wms_p90": 0.9674313822178536, "gt_coverage": 0.9699499165275458, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "c6155ec1d403bc84e1837b7b1c8d5f9052a606aceb0957e47bb3f4227fed358e", "metadata_sha256": "f3965c59ba7446c26837f265561095cabbb1f8e82f03e82d6dec5d42093c5411", "source_file": "long20_20260918/local/train/previews/18297a653ce83463b04bc637/reference.npz", "window_s": [ 0.06666666666666667, 20.033333333333335 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.761112836873094e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c60ff0de1eb5b392de8f6e3a40a1ac772910ef53135e2050899c60951b44bab4", "mano_sha256": "f85cfd0d3c6009fe131b19ba206d4e079bb451a1ff4e1b6800cffee63df74fdf", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "189a3d3042353aebbb9bb861", "sample_id": "42e13d6a38627f5eff7af4d3", "file": "20_seconds/ARCTIC/42e13d6a38627f5eff7af4d3_MANO_GT_20.00s.mp4", "recording_id": "s05/phone_use_02", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 18/600 output frames; may include genuine out-of-view frames", "id": "189a3d3042353aebbb9bb861", "aliases": [ "20_seconds/ARCTIC/64e8e0803235f1f1e65e7995_MANO_GT_20.00s.mp4" ], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/42e13d6a38627f5eff7af4d3_MANO_GT_20.00s.mp4", "bytes": 5628823, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05223867638213719, "p90": 0.1263320953795143, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 11.518358111861746, "p90": 30.861663679205734, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.05653832541114347, "p90": 0.28017348052891294, "valid_pairs": 418, "valid_duration_s": 13.93333333333333, "coverage": 0.697829716193656 }, "cmh_m_s": { "median": 0.05265119957976461, "p90": 0.12590772095191446, "valid_pairs": 418, "valid_duration_s": 13.93333333333333, "coverage": 0.697829716193656 }, "rhcm_continuous_estimate_m_s": { "median": 0.08410663908033253, "p90": 0.2761493281872902, "valid_pairs": 418, "valid_duration_s": 13.93333333333333, "coverage": 0.697829716193656 }, "rhcm_discrete_m_s": { "median": 0.08410183642520414, "p90": 0.2744897685405857, "valid_pairs": 418, "valid_duration_s": 13.93333333333333, "coverage": 0.697829716193656 }, "wms": { "median": 0.7361628950818083, "p90": 0.94093507445483, "valid_pairs": 415, "valid_duration_s": 13.83333333333333, "coverage": 0.692821368948247 }, "rham_deg_s": { "median": 45.090414863172455, "p90": 141.6342974281466, "valid_pairs": 418, "valid_duration_s": 13.93333333333333, "coverage": 0.697829716193656 }, "speed_spearman": 0.3193801469541385, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08090521469973687, "p90": 0.4100318386571634, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652754 }, "cmh_m_s": { "median": 0.04627896515124863, "p90": 0.11889724988928141, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652754 }, "rhcm_continuous_estimate_m_s": { "median": 0.09440119465942277, "p90": 0.4036071702502688, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652754 }, "rhcm_discrete_m_s": { "median": 0.09418730749523481, "p90": 0.40506140868336765, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652754 }, "wms": { "median": 0.7593435976605709, "p90": 0.9484816475655917, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652754 }, "rham_deg_s": { "median": 56.03387556537981, "p90": 181.40286339144495, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652754 }, "speed_spearman": 0.3492604055680193, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1263320953795143, "camera_angular_speed_p90_deg_s": 30.861663679205734, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4100318386571634, "hand_camera_speed_p90_m_s": 0.40506140868336765, "hand_world_valid_fraction": 0.7675, "hand_camera_valid_fraction": 0.7675, "cmh_p90_m_s": 0.12590772095191446, "rhcm_p90_m_s": 0.40506140868336765, "rham_p90_deg_s": 181.40286339144495, "wms_p90": 0.9484816475655917, "gt_coverage": 0.8196994991652754, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "96254783402e6e711186c7491ca1fdb562f36f85f2c389d405c5d55c86225132", "metadata_sha256": "561f6bf58cf64da92881c4f8675003cdaa606fdc5993f96c97328d9a4d294a93", "source_file": "long20_20260918/local/benchmark/previews/189a3d3042353aebbb9bb861/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.579669976578771e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e00f93546c9f1d32c1e4634ca46113b703c8a9df251846806d91d95f9ab44075", "mano_sha256": "297dda1426aaf587e5946c37396a02e74640a1761d3a02c12d435042ab312186", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "18c6d5c9116204aff0c86dea", "sample_id": "586202a0f40fed8c18869081", "file": "20_seconds/ARCTIC/586202a0f40fed8c18869081_MANO_GT_20.00s.mp4", "recording_id": "s02/microwave_grab_01", "source_first_frame": "85", "source_last_frame": "684", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 84/600 output frames; may include genuine out-of-view frames", "id": "18c6d5c9116204aff0c86dea", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/586202a0f40fed8c18869081_MANO_GT_20.00s.mp4", "bytes": 4533868, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02924995799697073, "p90": 0.08235458861055346, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 7.320109334963387, "p90": 15.379825706249038, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.16189872076958112, "p90": 0.42328077793328717, "valid_pairs": 455, "valid_duration_s": 15.166666666666668, "coverage": 0.7595993322203672 }, "cmh_m_s": { "median": 0.032173338703596704, "p90": 0.06911115611125448, "valid_pairs": 455, "valid_duration_s": 15.166666666666668, "coverage": 0.7595993322203672 }, "rhcm_continuous_estimate_m_s": { "median": 0.15105442826731869, "p90": 0.43570031188660907, "valid_pairs": 455, "valid_duration_s": 15.166666666666668, "coverage": 0.7595993322203672 }, "rhcm_discrete_m_s": { "median": 0.15145815664273093, "p90": 0.4356938587129081, "valid_pairs": 455, "valid_duration_s": 15.166666666666668, "coverage": 0.7595993322203672 }, "wms": { "median": 0.8328989472117212, "p90": 0.9696595583438495, "valid_pairs": 455, "valid_duration_s": 15.166666666666668, "coverage": 0.7595993322203672 }, "rham_deg_s": { "median": 23.93336273179356, "p90": 78.69022322566238, "valid_pairs": 455, "valid_duration_s": 15.166666666666668, "coverage": 0.7595993322203672 }, "speed_spearman": 0.1626312477228313, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.23193346402369672, "p90": 0.5795825615736829, "valid_pairs": 513, "valid_duration_s": 17.099999999999998, "coverage": 0.8564273789649414 }, "cmh_m_s": { "median": 0.052634718763838705, "p90": 0.11516618878110273, "valid_pairs": 513, "valid_duration_s": 17.099999999999998, "coverage": 0.8564273789649414 }, "rhcm_continuous_estimate_m_s": { "median": 0.1874951770704509, "p90": 0.531603966303303, "valid_pairs": 513, "valid_duration_s": 17.099999999999998, "coverage": 0.8564273789649414 }, "rhcm_discrete_m_s": { "median": 0.18744384621408297, "p90": 0.531483133726693, "valid_pairs": 513, "valid_duration_s": 17.099999999999998, "coverage": 0.8564273789649414 }, "wms": { "median": 0.7047772792834811, "p90": 0.9115587575145115, "valid_pairs": 513, "valid_duration_s": 17.099999999999998, "coverage": 0.8564273789649414 }, "rham_deg_s": { "median": 28.79349811636566, "p90": 82.1529766106362, "valid_pairs": 513, "valid_duration_s": 17.099999999999998, "coverage": 0.8564273789649414 }, "speed_spearman": 0.39925693520035505, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.08235458861055346, "camera_angular_speed_p90_deg_s": 15.379825706249038, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5795825615736829, "hand_camera_speed_p90_m_s": 0.531483133726693, "hand_world_valid_fraction": 0.8133333333333334, "hand_camera_valid_fraction": 0.8133333333333334, "cmh_p90_m_s": 0.11516618878110273, "rhcm_p90_m_s": 0.531483133726693, "rham_p90_deg_s": 82.1529766106362, "wms_p90": 0.9696595583438495, "gt_coverage": 0.8564273789649414, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "8155caf147416a30b125903d2706bfaafde76b0a9dea8ef02e4e4f7c026f8961", "metadata_sha256": "0e6521989d846155232aadc4c9a71faba7500c3db315281940b3dcea98055c55", "source_file": "long20_20260918/local/train/previews/18c6d5c9116204aff0c86dea/reference.npz", "window_s": [ 2.7666666666666666, 22.733333333333334 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.528924144620228e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "64a88ee48a40a31536213704c703c449f97904821679533db73836b2e5e71a38", "mano_sha256": "a36cb099120a045e4ff569c73081f5463ebe97333b079dd1e1f067972062c94d", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "1acc0d78161d5e5d88e6d735", "sample_id": "e86a5cf581b4919168238951", "file": "20_seconds/ARCTIC/e86a5cf581b4919168238951_MANO_GT_20.00s.mp4", "recording_id": "s05/mixer_grab_01", "source_first_frame": "181", "source_last_frame": "780", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 58/600 output frames; may include genuine out-of-view frames", "id": "1acc0d78161d5e5d88e6d735", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/e86a5cf581b4919168238951_MANO_GT_20.00s.mp4", "bytes": 5367481, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.11607869132897873, "p90": 0.22290929777526866, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 16.30863434900578, "p90": 41.4660083814844, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.209097543053972, "p90": 0.5653259532754141, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.8714524207011688 }, "cmh_m_s": { "median": 0.1048031516272155, "p90": 0.2710851496887675, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.8714524207011688 }, "rhcm_continuous_estimate_m_s": { "median": 0.1968953298961706, "p90": 0.5000275267430961, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.8714524207011688 }, "rhcm_discrete_m_s": { "median": 0.19641682391812762, "p90": 0.4997741915953674, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.8714524207011688 }, "wms": { "median": 0.6691474909727427, "p90": 0.859967033359016, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.8714524207011688 }, "rham_deg_s": { "median": 53.9696308196534, "p90": 123.6430754836694, "valid_pairs": 522, "valid_duration_s": 17.400000000000002, "coverage": 0.8714524207011688 }, "speed_spearman": 0.5861850455628973, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.22307276084868022, "p90": 0.604163075870449, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "cmh_m_s": { "median": 0.13878667170977643, "p90": 0.29280537135263013, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "rhcm_continuous_estimate_m_s": { "median": 0.20970708941107907, "p90": 0.5489110366263986, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "rhcm_discrete_m_s": { "median": 0.2098122014179491, "p90": 0.5483349974975111, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "wms": { "median": 0.6032118409190321, "p90": 0.8928367165857147, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "rham_deg_s": { "median": 47.75022586146619, "p90": 147.27757821623803, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "speed_spearman": 0.6541793208589671, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.22290929777526866, "camera_angular_speed_p90_deg_s": 41.4660083814844, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.604163075870449, "hand_camera_speed_p90_m_s": 0.5483349974975111, "hand_world_valid_fraction": 0.8583333333333333, "hand_camera_valid_fraction": 0.8583333333333333, "cmh_p90_m_s": 0.29280537135263013, "rhcm_p90_m_s": 0.5483349974975111, "rham_p90_deg_s": 147.27757821623803, "wms_p90": 0.8928367165857147, "gt_coverage": 0.8714524207011688, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "ffec948891ab8918ed312f51baafeabfb9561f4b7c068626c2423dc9e0071d10", "metadata_sha256": "95befa3a2a0711b7ef8c7a8ba46ccc5bdee7f56a66dcd19d6bed0c8b90d08127", "source_file": "long20_20260918/local/benchmark/previews/1acc0d78161d5e5d88e6d735/reference.npz", "window_s": [ 6.0, 25.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.579669976578771e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c33474d5bff79015ab629815a298e5ffe84f2da146e39b368c0a6e3a63922280", "mano_sha256": "90cfd7fe025b7509e2bd613ed6d5b9d7d127ea04d0527c27a4e0489c06db7ace", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "1de6e0124a18b987f2a70400", "sample_id": "6ed5ac44bcb4d8f13f8ee50b", "file": "20_seconds/ARCTIC/6ed5ac44bcb4d8f13f8ee50b_MANO_GT_20.00s.mp4", "recording_id": "s05/phone_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 84/600 output frames; may include genuine out-of-view frames", "id": "1de6e0124a18b987f2a70400", "aliases": [ "20_seconds/ARCTIC/d73977c904c5e7f45b16d42a_MANO_GT_20.00s.mp4" ], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/6ed5ac44bcb4d8f13f8ee50b_MANO_GT_20.00s.mp4", "bytes": 5903417, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.07990298653438922, "p90": 0.195392540789649, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 17.115506359323344, "p90": 48.09945873668529, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.1441424357981887, "p90": 0.38874662220864503, "valid_pairs": 421, "valid_duration_s": 14.033333333333331, "coverage": 0.7028380634390651 }, "cmh_m_s": { "median": 0.08106465366892188, "p90": 0.22425382175951356, "valid_pairs": 421, "valid_duration_s": 14.033333333333331, "coverage": 0.7028380634390651 }, "rhcm_continuous_estimate_m_s": { "median": 0.12558098584542574, "p90": 0.4086756241007915, "valid_pairs": 421, "valid_duration_s": 14.033333333333331, "coverage": 0.7028380634390651 }, "rhcm_discrete_m_s": { "median": 0.12592633962122124, "p90": 0.40788564667618676, "valid_pairs": 421, "valid_duration_s": 14.033333333333331, "coverage": 0.7028380634390651 }, "wms": { "median": 0.664343358751725, "p90": 0.9303379785791047, "valid_pairs": 421, "valid_duration_s": 14.033333333333331, "coverage": 0.7028380634390651 }, "rham_deg_s": { "median": 75.06276452929235, "p90": 180.53691394524736, "valid_pairs": 421, "valid_duration_s": 14.033333333333331, "coverage": 0.7028380634390651 }, "speed_spearman": 0.44729574778431597, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.10756734368845464, "p90": 0.404359005880928, "valid_pairs": 439, "valid_duration_s": 14.633333333333331, "coverage": 0.7328881469115192 }, "cmh_m_s": { "median": 0.09916794774079608, "p90": 0.2634981636728277, "valid_pairs": 439, "valid_duration_s": 14.633333333333331, "coverage": 0.7328881469115192 }, "rhcm_continuous_estimate_m_s": { "median": 0.12137149250109071, "p90": 0.3129729415096184, "valid_pairs": 439, "valid_duration_s": 14.633333333333331, "coverage": 0.7328881469115192 }, "rhcm_discrete_m_s": { "median": 0.12130961936994343, "p90": 0.31311847777160506, "valid_pairs": 439, "valid_duration_s": 14.633333333333331, "coverage": 0.7328881469115192 }, "wms": { "median": 0.6518113301409559, "p90": 0.9201762709279933, "valid_pairs": 439, "valid_duration_s": 14.633333333333331, "coverage": 0.7328881469115192 }, "rham_deg_s": { "median": 64.60187392085352, "p90": 199.53423807341605, "valid_pairs": 439, "valid_duration_s": 14.633333333333331, "coverage": 0.7328881469115192 }, "speed_spearman": 0.5485175183040818, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.195392540789649, "camera_angular_speed_p90_deg_s": 48.09945873668529, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.404359005880928, "hand_camera_speed_p90_m_s": 0.40788564667618676, "hand_world_valid_fraction": 0.7308333333333333, "hand_camera_valid_fraction": 0.7308333333333333, "cmh_p90_m_s": 0.2634981636728277, "rhcm_p90_m_s": 0.40788564667618676, "rham_p90_deg_s": 199.53423807341605, "wms_p90": 0.9303379785791047, "gt_coverage": 0.7328881469115192, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "6eb996d4b8d08607ac72891b9d27981dcf0a746635096bdef0f29d8152e1204e", "metadata_sha256": "564488196a89a44ada8f1000818d27f5d259bbabe804001d07b179a21532a4e2", "source_file": "long20_20260918/local/benchmark/previews/1de6e0124a18b987f2a70400/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.458205583648681e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3bb3ceeef38cff1b6123e384d08b7058b6903e65a3b9ff09bea0311db7f823e9", "mano_sha256": "33aede2570c2c2bc68657893cc9585ce2da61a0cb225879d11815a2df151d091", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "1df4591a2455a70052e267d1", "sample_id": "7b5b1c0930ccc3a587f74ed5", "file": "20_seconds/ARCTIC/7b5b1c0930ccc3a587f74ed5_MANO_GT_20.00s.mp4", "recording_id": "s05/notebook_use_02", "source_first_frame": "130", "source_last_frame": "729", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 69/600 output frames; may include genuine out-of-view frames", "id": "1df4591a2455a70052e267d1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/7b5b1c0930ccc3a587f74ed5_MANO_GT_20.00s.mp4", "bytes": 6177919, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.09975217290542156, "p90": 0.2096076791223976, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 21.12064114780694, "p90": 46.82578772925703, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.11497260158186706, "p90": 0.3347524322580997, "valid_pairs": 479, "valid_duration_s": 15.966666666666663, "coverage": 0.7996661101836393 }, "cmh_m_s": { "median": 0.12071229367328051, "p90": 0.2796264035725593, "valid_pairs": 479, "valid_duration_s": 15.966666666666663, "coverage": 0.7996661101836393 }, "rhcm_continuous_estimate_m_s": { "median": 0.1950270646661975, "p90": 0.35952561407985967, "valid_pairs": 479, "valid_duration_s": 15.966666666666663, "coverage": 0.7996661101836393 }, "rhcm_discrete_m_s": { "median": 0.1948252963383895, "p90": 0.360540904395091, "valid_pairs": 479, "valid_duration_s": 15.966666666666663, "coverage": 0.7996661101836393 }, "wms": { "median": 0.7118743872555529, "p90": 0.8985837163614918, "valid_pairs": 479, "valid_duration_s": 15.966666666666663, "coverage": 0.7996661101836393 }, "rham_deg_s": { "median": 62.772902707127834, "p90": 200.7311069715938, "valid_pairs": 479, "valid_duration_s": 15.966666666666663, "coverage": 0.7996661101836393 }, "speed_spearman": 0.17427979315344905, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09999134725942614, "p90": 0.22047982721433781, "valid_pairs": 462, "valid_duration_s": 15.400000000000002, "coverage": 0.7712854757929885 }, "cmh_m_s": { "median": 0.1277669440867361, "p90": 0.2729820995006291, "valid_pairs": 462, "valid_duration_s": 15.400000000000002, "coverage": 0.7712854757929885 }, "rhcm_continuous_estimate_m_s": { "median": 0.1748191471798046, "p90": 0.3282122229126302, "valid_pairs": 462, "valid_duration_s": 15.400000000000002, "coverage": 0.7712854757929885 }, "rhcm_discrete_m_s": { "median": 0.17559016367970065, "p90": 0.32752959271124527, "valid_pairs": 462, "valid_duration_s": 15.400000000000002, "coverage": 0.7712854757929885 }, "wms": { "median": 0.717840606727458, "p90": 0.9277076967005241, "valid_pairs": 462, "valid_duration_s": 15.400000000000002, "coverage": 0.7712854757929885 }, "rham_deg_s": { "median": 59.606208544880474, "p90": 182.90355715417599, "valid_pairs": 462, "valid_duration_s": 15.400000000000002, "coverage": 0.7712854757929885 }, "speed_spearman": 0.27637750666849764, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.2096076791223976, "camera_angular_speed_p90_deg_s": 46.82578772925703, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3347524322580997, "hand_camera_speed_p90_m_s": 0.360540904395091, "hand_world_valid_fraction": 0.7925, "hand_camera_valid_fraction": 0.7925, "cmh_p90_m_s": 0.2796264035725593, "rhcm_p90_m_s": 0.360540904395091, "rham_p90_deg_s": 200.7311069715938, "wms_p90": 0.9277076967005241, "gt_coverage": 0.7996661101836393, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "79995934d68d8008fd65782afd48a5ce33c38d21b5721cfd4b69634e5e3a333a", "metadata_sha256": "102e5d452d272dd47ef89068336642c7ea536cf15091408ab5cfe720f2d46ec9", "source_file": "long20_20260918/local/benchmark/previews/1df4591a2455a70052e267d1/reference.npz", "window_s": [ 4.3, 24.266666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.484050470211362e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "177a3611510facf9759cced0aa5f4217c27aa261594075c8a9de844832bb8dc4", "mano_sha256": "7b31bee3e3039e2d34dd42fded0fbb71accccde208937dde35c76d6bee0b369b", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "False", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "204deb0a12168e908bb97387", "sample_id": "e5924361672fad11c357e0c1", "file": "20_seconds/ARCTIC/e5924361672fad11c357e0c1_MANO_GT_20.00s.mp4", "recording_id": "s10/box_use_01", "source_first_frame": "121", "source_last_frame": "720", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 60/600 output frames; may include genuine out-of-view frames", "id": "204deb0a12168e908bb97387", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/e5924361672fad11c357e0c1_MANO_GT_20.00s.mp4", "bytes": 4497402, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03396617511702949, "p90": 0.07577590077922407, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 5.0958160137194275, "p90": 10.307935552343096, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06127639160709986, "p90": 0.44412796950022027, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "cmh_m_s": { "median": 0.0374852974635395, "p90": 0.08273872164756556, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "rhcm_continuous_estimate_m_s": { "median": 0.07114647369577787, "p90": 0.4165401404645457, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "rhcm_discrete_m_s": { "median": 0.07113856119786971, "p90": 0.4165997436708287, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "wms": { "median": 0.7289658872633555, "p90": 0.9500748033219398, "valid_pairs": 536, "valid_duration_s": 17.866666666666667, "coverage": 0.8948247078464108 }, "rham_deg_s": { "median": 24.850475367921167, "p90": 124.30083851665289, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "speed_spearman": 0.4554273483641512, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.0446865258265373, "p90": 0.32720937422703633, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "cmh_m_s": { "median": 0.03949492629930913, "p90": 0.07661239779656064, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "rhcm_continuous_estimate_m_s": { "median": 0.07338618799785157, "p90": 0.32902910279012754, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "rhcm_discrete_m_s": { "median": 0.07343041391441675, "p90": 0.32900410469271746, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "wms": { "median": 0.7978179539660979, "p90": 0.9481039935929362, "valid_pairs": 534, "valid_duration_s": 17.8, "coverage": 0.8914858096828048 }, "rham_deg_s": { "median": 22.549136405639654, "p90": 94.46925350668998, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "speed_spearman": -0.12197084513377665, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.07577590077922407, "camera_angular_speed_p90_deg_s": 10.307935552343096, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.44412796950022027, "hand_camera_speed_p90_m_s": 0.4165997436708287, "hand_world_valid_fraction": 0.8975, "hand_camera_valid_fraction": 0.8975, "cmh_p90_m_s": 0.08273872164756556, "rhcm_p90_m_s": 0.4165997436708287, "rham_p90_deg_s": 124.30083851665289, "wms_p90": 0.9500748033219398, "gt_coverage": 0.8998330550918197, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "2b4b434f832583158ad8b69668445889f8a90d8565c2d66a0f09b8c264b5ae7b", "metadata_sha256": "d42b391fd6602952f70fe653337d409b8f7328b4c8cd9b6e08b84219d3012536", "source_file": "long20_20260918/local/train/previews/204deb0a12168e908bb97387/reference.npz", "window_s": [ 4.0, 23.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.685680459230494e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c874d6d7a7c5e42edc2325413de9e17e20cea3866a2e56504ac8157c02b6b45a", "mano_sha256": "90ec65ee52896a3a49e62facd14cbc76f183de7e0626b16533cce0fa2dd37fee", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "2139b5c5f8411f49faac49f5", "sample_id": "d5ec2039f676088356952f35", "file": "20_seconds/ARCTIC/d5ec2039f676088356952f35_MANO_GT_20.00s.mp4", "recording_id": "s06/laptop_use_01_retake", "source_first_frame": "1", "source_last_frame": "600", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 123/600 output frames; may include genuine out-of-view frames", "id": "2139b5c5f8411f49faac49f5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/d5ec2039f676088356952f35_MANO_GT_20.00s.mp4", "bytes": 5394700, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05318809429656001, "p90": 0.11026318483412727, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 10.254726110557932, "p90": 24.70218198421748, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07853693083443392, "p90": 0.32494043568965064, "valid_pairs": 474, "valid_duration_s": 15.799999999999997, "coverage": 0.7913188647746243 }, "cmh_m_s": { "median": 0.07114283423199533, "p90": 0.1848543325278386, "valid_pairs": 474, "valid_duration_s": 15.799999999999997, "coverage": 0.7913188647746243 }, "rhcm_continuous_estimate_m_s": { "median": 0.11673752853083436, "p90": 0.4417148857750556, "valid_pairs": 474, "valid_duration_s": 15.799999999999997, "coverage": 0.7913188647746243 }, "rhcm_discrete_m_s": { "median": 0.11679932551855846, "p90": 0.442000890465642, "valid_pairs": 474, "valid_duration_s": 15.799999999999997, "coverage": 0.7913188647746243 }, "wms": { "median": 0.8088398433819084, "p90": 0.9699744309035282, "valid_pairs": 472, "valid_duration_s": 15.73333333333333, "coverage": 0.7879799666110183 }, "rham_deg_s": { "median": 31.64810367670228, "p90": 133.85775698706172, "valid_pairs": 474, "valid_duration_s": 15.799999999999997, "coverage": 0.7913188647746243 }, "speed_spearman": 0.12218757614726196, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.20734209231207296, "p90": 0.6111732320488751, "valid_pairs": 404, "valid_duration_s": 13.466666666666665, "coverage": 0.674457429048414 }, "cmh_m_s": { "median": 0.06432178330137858, "p90": 0.14135949607668788, "valid_pairs": 404, "valid_duration_s": 13.466666666666665, "coverage": 0.674457429048414 }, "rhcm_continuous_estimate_m_s": { "median": 0.22991692049585216, "p90": 0.6302938621941875, "valid_pairs": 404, "valid_duration_s": 13.466666666666665, "coverage": 0.674457429048414 }, "rhcm_discrete_m_s": { "median": 0.23011692322244814, "p90": 0.630136645901038, "valid_pairs": 404, "valid_duration_s": 13.466666666666665, "coverage": 0.674457429048414 }, "wms": { "median": 0.8737637793079384, "p90": 0.9772659888486342, "valid_pairs": 402, "valid_duration_s": 13.399999999999999, "coverage": 0.671118530884808 }, "rham_deg_s": { "median": 103.02019220262086, "p90": 227.07509027907426, "valid_pairs": 404, "valid_duration_s": 13.466666666666665, "coverage": 0.674457429048414 }, "speed_spearman": 0.06746995625948037, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.11026318483412727, "camera_angular_speed_p90_deg_s": 24.70218198421748, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6111732320488751, "hand_camera_speed_p90_m_s": 0.630136645901038, "hand_world_valid_fraction": 0.7391666666666666, "hand_camera_valid_fraction": 0.7391666666666666, "cmh_p90_m_s": 0.1848543325278386, "rhcm_p90_m_s": 0.630136645901038, "rham_p90_deg_s": 227.07509027907426, "wms_p90": 0.9772659888486342, "gt_coverage": 0.7913188647746243, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "1b7c8e5a848d16627bfd3755058757c63e605700189bde113278513eb0b650ae", "metadata_sha256": "3e9014660fdd3e60cacde8492fecf29d3f2d97546712ccc27f41e7492186489f", "source_file": "long20_20260918/local/train/previews/2139b5c5f8411f49faac49f5/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.744904746134193e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "cd925d1b94f4992d26b0d4db63a8829e8a2a1be64620880037d6b119e4588118", "mano_sha256": "bc8df8bbe1ad4c25be66a917e0b472be7ba9e4ed2fb8dccedaeb6bbbfd6ae3ef", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "226d5978037578f3f1997370", "sample_id": "e632aa71904e157159c9d816", "file": "20_seconds/ARCTIC/e632aa71904e157159c9d816_MANO_GT_20.00s.mp4", "recording_id": "s05/notebook_grab_01", "source_first_frame": "76", "source_last_frame": "675", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 25/600 output frames; may include genuine out-of-view frames", "id": "226d5978037578f3f1997370", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/e632aa71904e157159c9d816_MANO_GT_20.00s.mp4", "bytes": 5605815, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.07457549272328635, "p90": 0.21017255982099753, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 17.74560673899451, "p90": 50.03338758317739, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.20774972171553385, "p90": 0.4845006057524175, "valid_pairs": 276, "valid_duration_s": 9.199999999999996, "coverage": 0.4607679465776292 }, "cmh_m_s": { "median": 0.08865687270925379, "p90": 0.2817581532380016, "valid_pairs": 276, "valid_duration_s": 9.199999999999996, "coverage": 0.4607679465776292 }, "rhcm_continuous_estimate_m_s": { "median": 0.24541464847769592, "p90": 0.4761221881916874, "valid_pairs": 276, "valid_duration_s": 9.199999999999996, "coverage": 0.4607679465776292 }, "rhcm_discrete_m_s": { "median": 0.24600632610125783, "p90": 0.4742186845694557, "valid_pairs": 276, "valid_duration_s": 9.199999999999996, "coverage": 0.4607679465776292 }, "wms": { "median": 0.7298701625701278, "p90": 0.9486941633830107, "valid_pairs": 276, "valid_duration_s": 9.199999999999996, "coverage": 0.4607679465776292 }, "rham_deg_s": { "median": 67.13644765175869, "p90": 208.0509065920284, "valid_pairs": 276, "valid_duration_s": 9.199999999999996, "coverage": 0.4607679465776292 }, "speed_spearman": 0.2476642741970006, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.21979592306448495, "p90": 0.556942513491525, "valid_pairs": 527, "valid_duration_s": 17.566666666666663, "coverage": 0.8797996661101836 }, "cmh_m_s": { "median": 0.11094243727992437, "p90": 0.244624487900289, "valid_pairs": 527, "valid_duration_s": 17.566666666666663, "coverage": 0.8797996661101836 }, "rhcm_continuous_estimate_m_s": { "median": 0.23272906966334567, "p90": 0.4769383663070591, "valid_pairs": 527, "valid_duration_s": 17.566666666666663, "coverage": 0.8797996661101836 }, "rhcm_discrete_m_s": { "median": 0.2327691067567386, "p90": 0.4775913434919446, "valid_pairs": 527, "valid_duration_s": 17.566666666666663, "coverage": 0.8797996661101836 }, "wms": { "median": 0.7085981607885554, "p90": 0.8997830019369312, "valid_pairs": 527, "valid_duration_s": 17.566666666666663, "coverage": 0.8797996661101836 }, "rham_deg_s": { "median": 83.94880968831859, "p90": 245.39939456623205, "valid_pairs": 527, "valid_duration_s": 17.566666666666663, "coverage": 0.8797996661101836 }, "speed_spearman": 0.30229858632792234, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.21017255982099753, "camera_angular_speed_p90_deg_s": 50.03338758317739, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.556942513491525, "hand_camera_speed_p90_m_s": 0.4775913434919446, "hand_world_valid_fraction": 0.6816666666666666, "hand_camera_valid_fraction": 0.6816666666666666, "cmh_p90_m_s": 0.2817581532380016, "rhcm_p90_m_s": 0.4775913434919446, "rham_p90_deg_s": 245.39939456623205, "wms_p90": 0.9486941633830107, "gt_coverage": 0.8797996661101836, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "ca30fbb9f3e149c2062601599bdf414dd5fc7b5d54a7496eba0fa47720dfe6ed", "metadata_sha256": "85f7f3a0b3769f98bb75148837b45a8de39fadc5e17b5a349f0f78b3c8180383", "source_file": "long20_20260918/local/benchmark/previews/226d5978037578f3f1997370/reference.npz", "window_s": [ 2.5, 22.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.671272446551088e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "25a36a36f7e297bd9d63f2116c888395092d5fba59d18327f4724ce1ed9abb8e", "mano_sha256": "0f5ab24acbb0f96b515b0fa7cf1e0a2292ce8daf0bdda8a8c64157030a6c3090", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "25e27387446e06b0e67f9a48", "sample_id": "04067654d510565edde6125b", "file": "20_seconds/ARCTIC/04067654d510565edde6125b_MANO_GT_20.00s.mp4", "recording_id": "s05/phone_use_03", "source_first_frame": "337", "source_last_frame": "936", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 48/600 output frames; may include genuine out-of-view frames", "id": "25e27387446e06b0e67f9a48", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/04067654d510565edde6125b_MANO_GT_20.00s.mp4", "bytes": 5424583, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.04803671573859868, "p90": 0.13904017264811808, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 10.817319035758361, "p90": 32.615470712053224, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08877850617758598, "p90": 0.21656467105060395, "valid_pairs": 551, "valid_duration_s": 18.366666666666667, "coverage": 0.9198664440734557 }, "cmh_m_s": { "median": 0.053055030915767215, "p90": 0.12094376543184181, "valid_pairs": 551, "valid_duration_s": 18.366666666666667, "coverage": 0.9198664440734557 }, "rhcm_continuous_estimate_m_s": { "median": 0.10000595457524313, "p90": 0.22202876875949423, "valid_pairs": 551, "valid_duration_s": 18.366666666666667, "coverage": 0.9198664440734557 }, "rhcm_discrete_m_s": { "median": 0.10132986250537687, "p90": 0.2218830503957739, "valid_pairs": 551, "valid_duration_s": 18.366666666666667, "coverage": 0.9198664440734557 }, "wms": { "median": 0.7457742205996614, "p90": 0.9445189769443744, "valid_pairs": 551, "valid_duration_s": 18.366666666666667, "coverage": 0.9198664440734557 }, "rham_deg_s": { "median": 62.330536032312885, "p90": 165.4874778887644, "valid_pairs": 551, "valid_duration_s": 18.366666666666667, "coverage": 0.9198664440734557 }, "speed_spearman": 0.3248502542986879, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.06493583561770731, "p90": 0.17110045407750485, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.7395659432387311 }, "cmh_m_s": { "median": 0.0591028842406366, "p90": 0.15466092815418028, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.7395659432387311 }, "rhcm_continuous_estimate_m_s": { "median": 0.09424512885087183, "p90": 0.19672694483736802, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.7395659432387311 }, "rhcm_discrete_m_s": { "median": 0.09401721102965799, "p90": 0.1975228769114406, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.7395659432387311 }, "wms": { "median": 0.7110757837895143, "p90": 0.9129637754015003, "valid_pairs": 442, "valid_duration_s": 14.733333333333333, "coverage": 0.7378964941569282 }, "rham_deg_s": { "median": 60.162618890505485, "p90": 195.45033692934393, "valid_pairs": 443, "valid_duration_s": 14.766666666666667, "coverage": 0.7395659432387311 }, "speed_spearman": 0.3042103725943854, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.13904017264811808, "camera_angular_speed_p90_deg_s": 32.615470712053224, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.21656467105060395, "hand_camera_speed_p90_m_s": 0.2218830503957739, "hand_world_valid_fraction": 0.8316666666666667, "hand_camera_valid_fraction": 0.8316666666666667, "cmh_p90_m_s": 0.15466092815418028, "rhcm_p90_m_s": 0.2218830503957739, "rham_p90_deg_s": 195.45033692934393, "wms_p90": 0.9445189769443744, "gt_coverage": 0.9198664440734557, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "bd890d1597d8f6435f2d1cce3af0c276950d315c65e3b14dd1733ebb4cbe54ba", "metadata_sha256": "7d51a54f2de1d06bdc6f22fbfaec314ef10018bcb0e048564a811712b3d8c3a1", "source_file": "long20_20260918/local/benchmark/previews/25e27387446e06b0e67f9a48/reference.npz", "window_s": [ 11.2, 31.166666666666668 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.552252871008668e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "05c5073428052725b7bbc67958c2569afc4f6d0c4470dcce9d5b0e319399569c", "mano_sha256": "0bc67323d7a82874c4708cdbcea32d9d005af93f700cdc1295c244031744e4f9", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "27249ee61a0b09c461f7428d", "sample_id": "d66267fe2ea0d0cef441592a", "file": "20_seconds/ARCTIC/d66267fe2ea0d0cef441592a_MANO_GT_20.00s.mp4", "recording_id": "s05/capsulemachine_use_01", "source_first_frame": "76", "source_last_frame": "675", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 44/600 output frames; may include genuine out-of-view frames", "id": "27249ee61a0b09c461f7428d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/d66267fe2ea0d0cef441592a_MANO_GT_20.00s.mp4", "bytes": 5344152, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0654768374009222, "p90": 0.14249045118776027, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 17.434144051130716, "p90": 42.35888006438791, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08919146989423649, "p90": 0.34189980339643555, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "cmh_m_s": { "median": 0.09394686088239221, "p90": 0.23692045016836286, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "rhcm_continuous_estimate_m_s": { "median": 0.15713035277464663, "p90": 0.38207556952064076, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "rhcm_discrete_m_s": { "median": 0.15711408814967592, "p90": 0.3814429966073777, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "wms": { "median": 0.8340742488230757, "p90": 0.9755843173439448, "valid_pairs": 552, "valid_duration_s": 18.4, "coverage": 0.9215358931552587 }, "rham_deg_s": { "median": 47.56182738315026, "p90": 139.41142458823995, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "speed_spearman": 0.1428523997077097, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.12415791421516296, "p90": 0.2960429554955532, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "cmh_m_s": { "median": 0.09038105848963719, "p90": 0.20742877880388358, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "rhcm_continuous_estimate_m_s": { "median": 0.15723944287682126, "p90": 0.38092475249492064, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "rhcm_discrete_m_s": { "median": 0.15731887462343178, "p90": 0.3813631483662296, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "wms": { "median": 0.8002774310134624, "p90": 0.9672582668478908, "valid_pairs": 554, "valid_duration_s": 18.46666666666667, "coverage": 0.924874791318865 }, "rham_deg_s": { "median": 51.951621807647676, "p90": 136.8569115545792, "valid_pairs": 555, "valid_duration_s": 18.5, "coverage": 0.9265442404006679 }, "speed_spearman": 0.12239516611903178, "hand_class": "low", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.14249045118776027, "camera_angular_speed_p90_deg_s": 42.35888006438791, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.34189980339643555, "hand_camera_speed_p90_m_s": 0.3814429966073777, "hand_world_valid_fraction": 0.9266666666666666, "hand_camera_valid_fraction": 0.9266666666666666, "cmh_p90_m_s": 0.23692045016836286, "rhcm_p90_m_s": 0.3814429966073777, "rham_p90_deg_s": 139.41142458823995, "wms_p90": 0.9755843173439448, "gt_coverage": 0.9265442404006679, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "49f68dfa982a9d658d8a2e70eac9496d385a2a61a296e850e632848fa60c607a", "metadata_sha256": "54bf5e7dfab189fb1881c94fcf0a65c583745776318e4e045091cf7be3db5216", "source_file": "long20_20260918/local/benchmark/previews/27249ee61a0b09c461f7428d/reference.npz", "window_s": [ 2.5, 22.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.585973723508303e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "21aeb6ef7472cd361a4258c40e8880a041c3680317e980c8c7eff76743b51e19", "mano_sha256": "904773ae4351fc2b523a084ec8bdc870a531a3b0f00233e0927419f9a88cb7a7", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "2ae5796c26145417ff360a79", "sample_id": "07658f26e558b29c670b6e50", "file": "20_seconds/ARCTIC/07658f26e558b29c670b6e50_MANO_GT_20.00s.mp4", "recording_id": "s07/capsulemachine_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 23/600 output frames; may include genuine out-of-view frames", "id": "2ae5796c26145417ff360a79", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/07658f26e558b29c670b6e50_MANO_GT_20.00s.mp4", "bytes": 4843874, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03379516506491417, "p90": 0.07407145130661573, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 5.84168355821163, "p90": 12.578508607298295, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07946850872196014, "p90": 0.24231874293892613, "valid_pairs": 576, "valid_duration_s": 19.2, "coverage": 0.9616026711185309 }, "cmh_m_s": { "median": 0.03472092244719337, "p90": 0.080225981362019, "valid_pairs": 576, "valid_duration_s": 19.2, "coverage": 0.9616026711185309 }, "rhcm_continuous_estimate_m_s": { "median": 0.08418083833698264, "p90": 0.2419564489686656, "valid_pairs": 576, "valid_duration_s": 19.2, "coverage": 0.9616026711185309 }, "rhcm_discrete_m_s": { "median": 0.08419514179599136, "p90": 0.24193215455443104, "valid_pairs": 576, "valid_duration_s": 19.2, "coverage": 0.9616026711185309 }, "wms": { "median": 0.8072778690880555, "p90": 0.9582659569614816, "valid_pairs": 575, "valid_duration_s": 19.166666666666664, "coverage": 0.9599332220367278 }, "rham_deg_s": { "median": 34.64407670861259, "p90": 104.26567699452053, "valid_pairs": 576, "valid_duration_s": 19.2, "coverage": 0.9616026711185309 }, "speed_spearman": 0.37681975736568457, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.08463937811414753, "p90": 0.2757338885395782, "valid_pairs": 513, "valid_duration_s": 17.099999999999994, "coverage": 0.8564273789649414 }, "cmh_m_s": { "median": 0.0357785131676353, "p90": 0.07519914095951903, "valid_pairs": 513, "valid_duration_s": 17.099999999999994, "coverage": 0.8564273789649414 }, "rhcm_continuous_estimate_m_s": { "median": 0.08285484947903865, "p90": 0.23043699121526265, "valid_pairs": 513, "valid_duration_s": 17.099999999999994, "coverage": 0.8564273789649414 }, "rhcm_discrete_m_s": { "median": 0.08279570994575924, "p90": 0.2304630526624894, "valid_pairs": 513, "valid_duration_s": 17.099999999999994, "coverage": 0.8564273789649414 }, "wms": { "median": 0.7046120494750655, "p90": 0.9248160335929562, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013355 }, "rham_deg_s": { "median": 36.45456267991844, "p90": 93.72402736936316, "valid_pairs": 513, "valid_duration_s": 17.099999999999994, "coverage": 0.8564273789649414 }, "speed_spearman": 0.43172982157485157, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.07407145130661573, "camera_angular_speed_p90_deg_s": 12.578508607298295, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2757338885395782, "hand_camera_speed_p90_m_s": 0.24193215455443104, "hand_world_valid_fraction": 0.9108333333333334, "hand_camera_valid_fraction": 0.9108333333333334, "cmh_p90_m_s": 0.080225981362019, "rhcm_p90_m_s": 0.24193215455443104, "rham_p90_deg_s": 104.26567699452053, "wms_p90": 0.9582659569614816, "gt_coverage": 0.9616026711185309, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "bdcef22b9f0dbf59f15b2409652c8fad9e8c5218f3b63c4c76b891eba1aa52a0", "metadata_sha256": "f91f2ca0ac448b66ac4cfa1ae46b946a2b94520f9b9f85e8df53b7a20c3ac90a", "source_file": "long20_20260918/local/train/previews/2ae5796c26145417ff360a79/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.621532547029659e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "eb402c915d6111cd869561925434a51f39d213b96ee34af2b1f5a737e651a98f", "mano_sha256": "11314f680ddcd426940b1815dc36f77ac8074a84c009262f108e3b07efb9274d", "duration_s": 19.634, "frames": 589, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "30ad3d4d8834821b8175cf1a", "sample_id": "e124444d54203442400d8b8b", "file": "short_originals_under_20s/ARCTIC/e124444d54203442400d8b8b_MANO_GT_19.63s.mp4", "recording_id": "s06/notebook_grab_01", "source_first_frame": "1", "source_last_frame": "589", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 180/589 output frames; may include genuine out-of-view frames", "id": "30ad3d4d8834821b8175cf1a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/ARCTIC/e124444d54203442400d8b8b_MANO_GT_19.63s.mp4", "bytes": 4506924, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.042306724314848165, "p90": 0.1785292490321799, "valid_pairs": 588, "valid_duration_s": 19.6, "coverage": 1.0 }, "camera_rotation": { "median": 7.7742332905341325, "p90": 31.677619306086264, "valid_pairs": 588, "valid_duration_s": 19.6, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.22716750070147046, "p90": 0.50200628974048, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5408163265306122 }, "cmh_m_s": { "median": 0.0703740953449328, "p90": 0.44412228511147106, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5408163265306122 }, "rhcm_continuous_estimate_m_s": { "median": 0.23220115740187597, "p90": 0.7405689936237937, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5408163265306122 }, "rhcm_discrete_m_s": { "median": 0.23218703933947668, "p90": 0.7402960387126273, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5408163265306122 }, "wms": { "median": 0.7873047373306138, "p90": 0.9719053989137166, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5408163265306122 }, "rham_deg_s": { "median": 81.84824447075613, "p90": 192.01907212778906, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5408163265306122 }, "speed_spearman": 0.22087187301696293, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.15966709438808757, "p90": 0.4414355736322862, "valid_pairs": 404, "valid_duration_s": 13.466666666666667, "coverage": 0.6870748299319728 }, "cmh_m_s": { "median": 0.0712220167043093, "p90": 0.27300365087066103, "valid_pairs": 404, "valid_duration_s": 13.466666666666667, "coverage": 0.6870748299319728 }, "rhcm_continuous_estimate_m_s": { "median": 0.18267064602097804, "p90": 0.5110170917349388, "valid_pairs": 404, "valid_duration_s": 13.466666666666667, "coverage": 0.6870748299319728 }, "rhcm_discrete_m_s": { "median": 0.18267989272278184, "p90": 0.5109713579555026, "valid_pairs": 404, "valid_duration_s": 13.466666666666667, "coverage": 0.6870748299319728 }, "wms": { "median": 0.7723765236750836, "p90": 0.9587106182807076, "valid_pairs": 403, "valid_duration_s": 13.433333333333334, "coverage": 0.6853741496598639 }, "rham_deg_s": { "median": 61.388397621842245, "p90": 151.58248844250338, "valid_pairs": 404, "valid_duration_s": 13.466666666666667, "coverage": 0.6870748299319728 }, "speed_spearman": 0.14598669130767503, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1785292490321799, "camera_angular_speed_p90_deg_s": 31.677619306086264, "camera_valid_fraction": 1.0, "camera_pairs": 588, "hand_world_speed_p90_m_s": 0.50200628974048, "hand_camera_speed_p90_m_s": 0.7402960387126273, "hand_world_valid_fraction": 0.6179966044142614, "hand_camera_valid_fraction": 0.6179966044142614, "cmh_p90_m_s": 0.44412228511147106, "rhcm_p90_m_s": 0.7402960387126273, "rham_p90_deg_s": 192.01907212778906, "wms_p90": 0.9719053989137166, "gt_coverage": 0.6870748299319728, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "d332cf33709c0c9bf5cfcb506179306dae9a0a5972ceb80f2b42c7ef63ba6570", "metadata_sha256": "593c4df3dd1194b173b2196552d5047c3af1e1b5dae78a5281b5ec7394291e00", "source_file": "long20_20260918/local/train/previews/30ad3d4d8834821b8175cf1a/reference.npz", "window_s": [ 0.0, 19.6 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 589, "common_world_residual_max_m": 4.878984946956307e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "9f9793885e121da71ba58200131b44092981ff61351eb64fe6457d2d50f106fc", "mano_sha256": "011ed32864533430f5477bc179518edaf9e6c2cdfd145bac0faa247bb28db675", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "31fc8ebac37428d432cf44f2", "sample_id": "54c8d77dd4130f2a15614ad7", "file": "20_seconds/ARCTIC/54c8d77dd4130f2a15614ad7_MANO_GT_20.00s.mp4", "recording_id": "s02/notebook_use_01", "source_first_frame": "111", "source_last_frame": "710", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 60/600 output frames; may include genuine out-of-view frames", "id": "31fc8ebac37428d432cf44f2", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/54c8d77dd4130f2a15614ad7_MANO_GT_20.00s.mp4", "bytes": 5299835, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.022503293438198162, "p90": 0.06323660020424492, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.8925827244309055, "p90": 7.274535852300923, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.032538357827956506, "p90": 0.21779838051278502, "valid_pairs": 537, "valid_duration_s": 17.900000000000002, "coverage": 0.8964941569282137 }, "cmh_m_s": { "median": 0.02662047905999568, "p90": 0.04543666937961615, "valid_pairs": 537, "valid_duration_s": 17.900000000000002, "coverage": 0.8964941569282137 }, "rhcm_continuous_estimate_m_s": { "median": 0.039145334145988746, "p90": 0.19256234075537004, "valid_pairs": 537, "valid_duration_s": 17.900000000000002, "coverage": 0.8964941569282137 }, "rhcm_discrete_m_s": { "median": 0.039142699484024934, "p90": 0.19267102486188453, "valid_pairs": 537, "valid_duration_s": 17.900000000000002, "coverage": 0.8964941569282137 }, "wms": { "median": 0.6457457581616017, "p90": 0.9044397317206584, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.8881469115191987 }, "rham_deg_s": { "median": 13.724802580505177, "p90": 73.64161088827498, "valid_pairs": 537, "valid_duration_s": 17.900000000000002, "coverage": 0.8964941569282137 }, "speed_spearman": 0.22425573840006421, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.19234434066991035, "p90": 0.39188207867696473, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "cmh_m_s": { "median": 0.023156200576848758, "p90": 0.0427160817397887, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "rhcm_continuous_estimate_m_s": { "median": 0.1893376904460032, "p90": 0.40161666430008497, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "rhcm_discrete_m_s": { "median": 0.18935023145918695, "p90": 0.40161564722211, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "wms": { "median": 0.9175034138052464, "p90": 0.9880230250927999, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "rham_deg_s": { "median": 60.0124944702459, "p90": 117.89927748365115, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "speed_spearman": 0.03555630656002402, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06323660020424492, "camera_angular_speed_p90_deg_s": 7.274535852300923, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.39188207867696473, "hand_camera_speed_p90_m_s": 0.40161564722211, "hand_world_valid_fraction": 0.8983333333333333, "hand_camera_valid_fraction": 0.8983333333333333, "cmh_p90_m_s": 0.04543666937961615, "rhcm_p90_m_s": 0.40161564722211, "rham_p90_deg_s": 117.89927748365115, "wms_p90": 0.9880230250927999, "gt_coverage": 0.8998330550918197, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "9a2a8ba09c94c91f1a98e91051fc6c27c7d087ae64a992352488835fd7979186", "metadata_sha256": "c6f2110b00fc470eed097bbd52f2df339f2556661885e71d00b52f335dd4fd60", "source_file": "long20_20260918/local/train/previews/31fc8ebac37428d432cf44f2/reference.npz", "window_s": [ 3.6333333333333333, 23.6 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.660953740672398e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6b9874685187a1313a017a4e4a55a3a61b92413cd66f68a239d399faa7b98091", "mano_sha256": "07cc9987e39651e2c2aca0bc54c8b5b2992bdb6e41e043438f85af0a31594faf", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "338ec7ff8233118451634b65", "sample_id": "aee0601977b622d1e35a5d8e", "file": "20_seconds/ARCTIC/aee0601977b622d1e35a5d8e_MANO_GT_20.00s.mp4", "recording_id": "s05/ketchup_grab_01", "source_first_frame": "169", "source_last_frame": "768", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 63/600 output frames; may include genuine out-of-view frames", "id": "338ec7ff8233118451634b65", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/aee0601977b622d1e35a5d8e_MANO_GT_20.00s.mp4", "bytes": 5435376, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08230803278041533, "p90": 0.2395397313657876, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 16.50055390824097, "p90": 50.66507389153733, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.24325144719816272, "p90": 0.7174086157628612, "valid_pairs": 388, "valid_duration_s": 12.933333333333334, "coverage": 0.6477462437395659 }, "cmh_m_s": { "median": 0.08941277636711134, "p90": 0.2754420555744865, "valid_pairs": 388, "valid_duration_s": 12.933333333333334, "coverage": 0.6477462437395659 }, "rhcm_continuous_estimate_m_s": { "median": 0.22173294186751707, "p90": 0.6000852222094631, "valid_pairs": 388, "valid_duration_s": 12.933333333333334, "coverage": 0.6477462437395659 }, "rhcm_discrete_m_s": { "median": 0.22145566474113143, "p90": 0.6002242643174086, "valid_pairs": 388, "valid_duration_s": 12.933333333333334, "coverage": 0.6477462437395659 }, "wms": { "median": 0.6991276649148873, "p90": 0.9272420690162357, "valid_pairs": 388, "valid_duration_s": 12.933333333333334, "coverage": 0.6477462437395659 }, "rham_deg_s": { "median": 83.77003584049176, "p90": 199.94047616632443, "valid_pairs": 388, "valid_duration_s": 12.933333333333334, "coverage": 0.6477462437395659 }, "speed_spearman": 0.3537043325840868, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.2873124568025025, "p90": 0.6461381341516378, "valid_pairs": 381, "valid_duration_s": 12.7, "coverage": 0.6360601001669448 }, "cmh_m_s": { "median": 0.10390703634018898, "p90": 0.26217347334308116, "valid_pairs": 381, "valid_duration_s": 12.7, "coverage": 0.6360601001669448 }, "rhcm_continuous_estimate_m_s": { "median": 0.25704419413562246, "p90": 0.5968315965410314, "valid_pairs": 381, "valid_duration_s": 12.7, "coverage": 0.6360601001669448 }, "rhcm_discrete_m_s": { "median": 0.25843562825271793, "p90": 0.5970021570773436, "valid_pairs": 381, "valid_duration_s": 12.7, "coverage": 0.6360601001669448 }, "wms": { "median": 0.7199641955114487, "p90": 0.9471083519605976, "valid_pairs": 380, "valid_duration_s": 12.666666666666666, "coverage": 0.6343906510851418 }, "rham_deg_s": { "median": 96.94109173616268, "p90": 270.1018040782166, "valid_pairs": 381, "valid_duration_s": 12.7, "coverage": 0.6360601001669448 }, "speed_spearman": 0.42753996133508937, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.2395397313657876, "camera_angular_speed_p90_deg_s": 50.66507389153733, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.7174086157628612, "hand_camera_speed_p90_m_s": 0.6002242643174086, "hand_world_valid_fraction": 0.6516666666666666, "hand_camera_valid_fraction": 0.6516666666666666, "cmh_p90_m_s": 0.2754420555744865, "rhcm_p90_m_s": 0.6002242643174086, "rham_p90_deg_s": 270.1018040782166, "wms_p90": 0.9471083519605976, "gt_coverage": 0.6477462437395659, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "444194679275bc17b4f044de25ae03c7d728554d59aaf486bc52fe88bc4f4d1c", "metadata_sha256": "e981d71c7153f3c207c8d0b83ddec4de9cdf7cf1639d576ffc3f9d4a0f014989", "source_file": "long20_20260918/local/benchmark/previews/338ec7ff8233118451634b65/reference.npz", "window_s": [ 5.6, 25.566666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.611102534756203e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "96929ef8a0e0da2ac10d9928b4bd688abd39fab83e79b7989e8ac34541d8a470", "mano_sha256": "0ac7fb2a70d38eb69a5db38fba41ba264e359efb0084dc4bd25b80e93acb7b2f", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "3dcc42498ec815a641037918", "sample_id": "483675fdbb5d740054c5300a", "file": "20_seconds/ARCTIC/483675fdbb5d740054c5300a_MANO_GT_20.00s.mp4", "recording_id": "s07/box_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 28/600 output frames; may include genuine out-of-view frames", "id": "3dcc42498ec815a641037918", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/483675fdbb5d740054c5300a_MANO_GT_20.00s.mp4", "bytes": 5297595, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.059698635040706716, "p90": 0.1384061532155031, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 9.267778146661527, "p90": 23.180237312322962, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07788360136000386, "p90": 0.34965992125819434, "valid_pairs": 571, "valid_duration_s": 19.03333333333333, "coverage": 0.9532554257095158 }, "cmh_m_s": { "median": 0.05885421416523216, "p90": 0.16220665387650735, "valid_pairs": 571, "valid_duration_s": 19.03333333333333, "coverage": 0.9532554257095158 }, "rhcm_continuous_estimate_m_s": { "median": 0.0730668269546721, "p90": 0.2751629521872924, "valid_pairs": 571, "valid_duration_s": 19.03333333333333, "coverage": 0.9532554257095158 }, "rhcm_discrete_m_s": { "median": 0.07324106846221144, "p90": 0.27514572450426095, "valid_pairs": 571, "valid_duration_s": 19.03333333333333, "coverage": 0.9532554257095158 }, "wms": { "median": 0.6279983179598237, "p90": 0.9425420336011351, "valid_pairs": 570, "valid_duration_s": 18.999999999999996, "coverage": 0.9515859766277127 }, "rham_deg_s": { "median": 22.978122445330587, "p90": 74.01484993395891, "valid_pairs": 571, "valid_duration_s": 19.03333333333333, "coverage": 0.9532554257095158 }, "speed_spearman": 0.38628032489217845, "hand_class": "high", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.2061891286800931, "p90": 0.4435328930558301, "valid_pairs": 163, "valid_duration_s": 5.433333333333333, "coverage": 0.2721202003338898 }, "cmh_m_s": { "median": 0.057201323761972046, "p90": 0.11684976031248528, "valid_pairs": 163, "valid_duration_s": 5.433333333333333, "coverage": 0.2721202003338898 }, "rhcm_continuous_estimate_m_s": { "median": 0.19021457869055516, "p90": 0.4013856842616882, "valid_pairs": 163, "valid_duration_s": 5.433333333333333, "coverage": 0.2721202003338898 }, "rhcm_discrete_m_s": { "median": 0.19013614739478737, "p90": 0.401401804992698, "valid_pairs": 163, "valid_duration_s": 5.433333333333333, "coverage": 0.2721202003338898 }, "wms": { "median": 0.7510921383346323, "p90": 0.911866724173473, "valid_pairs": 163, "valid_duration_s": 5.433333333333333, "coverage": 0.2721202003338898 }, "rham_deg_s": { "median": 36.536043069715866, "p90": 129.96583368997673, "valid_pairs": 163, "valid_duration_s": 5.433333333333333, "coverage": 0.2721202003338898 }, "speed_spearman": 0.29075154759727556, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1384061532155031, "camera_angular_speed_p90_deg_s": 23.180237312322962, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4435328930558301, "hand_camera_speed_p90_m_s": 0.401401804992698, "hand_world_valid_fraction": 0.6191666666666666, "hand_camera_valid_fraction": 0.6191666666666666, "cmh_p90_m_s": 0.16220665387650735, "rhcm_p90_m_s": 0.401401804992698, "rham_p90_deg_s": 129.96583368997673, "wms_p90": 0.9425420336011351, "gt_coverage": 0.9532554257095158, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "72dbd21fcd7401137b5be98730e285d6f2554cc8cf435bebac423ba34783fdc9", "metadata_sha256": "46cce714234e63131fc043c25d6e61e8132856c6bf82af8661cc3af6e84a997e", "source_file": "long20_20260918/local/train/previews/3dcc42498ec815a641037918/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.611102534756203e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3a67784997978521499d46295bcb3363f2408c4a5bdb7d0f5bc57a9bb2d846e3", "mano_sha256": "109d623c1989a4a95df7538b34986994ad06da765e624c6e25faf8abedcfca10", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "3eaa0f9460677918a8747333", "sample_id": "caa7e059668c5305774948dd", "file": "20_seconds/ARCTIC/caa7e059668c5305774948dd_MANO_GT_20.00s.mp4", "recording_id": "s05/microwave_use_01", "source_first_frame": "178", "source_last_frame": "777", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 67/600 output frames; may include genuine out-of-view frames", "id": "3eaa0f9460677918a8747333", "aliases": [ "20_seconds/ARCTIC/fff9c4fc04eee1fd8f44a939_MANO_GT_20.00s.mp4" ], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/caa7e059668c5305774948dd_MANO_GT_20.00s.mp4", "bytes": 5024171, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.10106224795235463, "p90": 0.18797291570324126, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 15.893091863028678, "p90": 35.576438294589614, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.12076680679911343, "p90": 0.46994192003586466, "valid_pairs": 531, "valid_duration_s": 17.7, "coverage": 0.8864774624373956 }, "cmh_m_s": { "median": 0.1005286898944789, "p90": 0.2530488131957799, "valid_pairs": 531, "valid_duration_s": 17.7, "coverage": 0.8864774624373956 }, "rhcm_continuous_estimate_m_s": { "median": 0.16519524228495558, "p90": 0.5209587218470874, "valid_pairs": 531, "valid_duration_s": 17.7, "coverage": 0.8864774624373956 }, "rhcm_discrete_m_s": { "median": 0.1651773147429426, "p90": 0.5200501306391921, "valid_pairs": 531, "valid_duration_s": 17.7, "coverage": 0.8864774624373956 }, "wms": { "median": 0.7818872350480088, "p90": 0.9457962323776041, "valid_pairs": 530, "valid_duration_s": 17.666666666666664, "coverage": 0.8848080133555924 }, "rham_deg_s": { "median": 52.70510307093539, "p90": 174.62142286711602, "valid_pairs": 531, "valid_duration_s": 17.7, "coverage": 0.8864774624373956 }, "speed_spearman": 0.3879823479389231, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.05504443468065585, "p90": 0.4240590198116527, "valid_pairs": 361, "valid_duration_s": 12.033333333333331, "coverage": 0.6026711185308846 }, "cmh_m_s": { "median": 0.1370960352758607, "p90": 0.2766823616982123, "valid_pairs": 361, "valid_duration_s": 12.033333333333331, "coverage": 0.6026711185308846 }, "rhcm_continuous_estimate_m_s": { "median": 0.1905054430760874, "p90": 0.4609592627027729, "valid_pairs": 361, "valid_duration_s": 12.033333333333331, "coverage": 0.6026711185308846 }, "rhcm_discrete_m_s": { "median": 0.1909512633027346, "p90": 0.46083163478851497, "valid_pairs": 361, "valid_duration_s": 12.033333333333331, "coverage": 0.6026711185308846 }, "wms": { "median": 0.8452153553668406, "p90": 0.9675282011220354, "valid_pairs": 361, "valid_duration_s": 12.033333333333331, "coverage": 0.6026711185308846 }, "rham_deg_s": { "median": 37.86423351243181, "p90": 223.1501813933862, "valid_pairs": 361, "valid_duration_s": 12.033333333333331, "coverage": 0.6026711185308846 }, "speed_spearman": -0.12014380965498946, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.18797291570324126, "camera_angular_speed_p90_deg_s": 35.576438294589614, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.46994192003586466, "hand_camera_speed_p90_m_s": 0.5200501306391921, "hand_world_valid_fraction": 0.7475, "hand_camera_valid_fraction": 0.7475, "cmh_p90_m_s": 0.2766823616982123, "rhcm_p90_m_s": 0.5200501306391921, "rham_p90_deg_s": 223.1501813933862, "wms_p90": 0.9675282011220354, "gt_coverage": 0.8864774624373956, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "1074a51de3b07c94fa180a060620f0c8b18e0d256fa755891dc8c6d04989ead2", "metadata_sha256": "268ff9daeb5ece1df990d29973267c3859120f7ca16d1310b2288f62b3591695", "source_file": "long20_20260918/local/benchmark/previews/3eaa0f9460677918a8747333/reference.npz", "window_s": [ 5.9, 25.866666666666667 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.509747244882934e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "03df3952b408228cd38ba063ad172b472a626b3d7cfa35449c86312b868c7abd", "mano_sha256": "c1cc2bd954295607991e93b91a7c2270fe29d0ed3c3f46ac9c83c5da2677b429", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "40d530cf67b9ffebfa41d349", "sample_id": "1a941564bcb391f9cf67594f", "file": "20_seconds/ARCTIC/1a941564bcb391f9cf67594f_MANO_GT_20.00s.mp4", "recording_id": "s07/laptop_use_02", "source_first_frame": "156", "source_last_frame": "755", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 79/600 output frames; may include genuine out-of-view frames", "id": "40d530cf67b9ffebfa41d349", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/1a941564bcb391f9cf67594f_MANO_GT_20.00s.mp4", "bytes": 5091837, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0524130271927875, "p90": 0.0944169827870631, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 11.280268314516265, "p90": 22.559627502856486, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.12439852351255445, "p90": 0.2793040490826713, "valid_pairs": 518, "valid_duration_s": 17.266666666666666, "coverage": 0.8647746243739566 }, "cmh_m_s": { "median": 0.0735879762387712, "p90": 0.14686367860509017, "valid_pairs": 518, "valid_duration_s": 17.266666666666666, "coverage": 0.8647746243739566 }, "rhcm_continuous_estimate_m_s": { "median": 0.07391118804254067, "p90": 0.22025337768420225, "valid_pairs": 518, "valid_duration_s": 17.266666666666666, "coverage": 0.8647746243739566 }, "rhcm_discrete_m_s": { "median": 0.07389872030288869, "p90": 0.22022380532938235, "valid_pairs": 518, "valid_duration_s": 17.266666666666666, "coverage": 0.8647746243739566 }, "wms": { "median": 0.4798699711723629, "p90": 0.8944447452792936, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "rham_deg_s": { "median": 30.39882157764184, "p90": 136.7695976651906, "valid_pairs": 518, "valid_duration_s": 17.266666666666666, "coverage": 0.8647746243739566 }, "speed_spearman": 0.5429080122229797, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.16572595597170522, "p90": 0.3743889269642294, "valid_pairs": 308, "valid_duration_s": 10.266666666666666, "coverage": 0.5141903171953255 }, "cmh_m_s": { "median": 0.05214059331004352, "p90": 0.1338171455662653, "valid_pairs": 308, "valid_duration_s": 10.266666666666666, "coverage": 0.5141903171953255 }, "rhcm_continuous_estimate_m_s": { "median": 0.12617852032142968, "p90": 0.3621645121871483, "valid_pairs": 308, "valid_duration_s": 10.266666666666666, "coverage": 0.5141903171953255 }, "rhcm_discrete_m_s": { "median": 0.1261485786023841, "p90": 0.362152674486223, "valid_pairs": 308, "valid_duration_s": 10.266666666666666, "coverage": 0.5141903171953255 }, "wms": { "median": 0.6898201236935889, "p90": 0.9242716761619761, "valid_pairs": 308, "valid_duration_s": 10.266666666666666, "coverage": 0.5141903171953255 }, "rham_deg_s": { "median": 30.537562029116142, "p90": 70.65001902553233, "valid_pairs": 308, "valid_duration_s": 10.266666666666666, "coverage": 0.5141903171953255 }, "speed_spearman": 0.25381578985196834, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0944169827870631, "camera_angular_speed_p90_deg_s": 22.559627502856486, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3743889269642294, "hand_camera_speed_p90_m_s": 0.362152674486223, "hand_world_valid_fraction": 0.6966666666666667, "hand_camera_valid_fraction": 0.6966666666666667, "cmh_p90_m_s": 0.14686367860509017, "rhcm_p90_m_s": 0.362152674486223, "rham_p90_deg_s": 136.7695976651906, "wms_p90": 0.9242716761619761, "gt_coverage": 0.8647746243739566, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "0fde2d07f9d3e7669fa7228157a5dfceefa3f7975d2b15bcf57ac2b73de77d3e", "metadata_sha256": "50263e3f148d4b7eca1ae5ae3d7b77ca1f665aaea93651dd17056fe1b8e7a278", "source_file": "long20_20260918/local/train/previews/40d530cf67b9ffebfa41d349/reference.npz", "window_s": [ 5.166666666666667, 25.133333333333333 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.972820174300706e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "818532a6fae781e7910f81daacea46fb98eabce36c7274a45986c81f6de22d1f", "mano_sha256": "930dadbe2c2b36100ae3c8bb794ef4751eb1eac90706227374990cd11fd2d3f7", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "46d68e78e2b44ae5a011094e", "sample_id": "e9d83fc7b7d5454d3909f051", "file": "20_seconds/ARCTIC/e9d83fc7b7d5454d3909f051_MANO_GT_20.00s.mp4", "recording_id": "s02/laptop_grab_01", "source_first_frame": "68", "source_last_frame": "667", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 75/600 output frames; may include genuine out-of-view frames", "id": "46d68e78e2b44ae5a011094e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/e9d83fc7b7d5454d3909f051_MANO_GT_20.00s.mp4", "bytes": 4822794, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.023438165050615318, "p90": 0.057376760842696134, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 6.286648528181564, "p90": 14.975154036443177, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.16895490091779183, "p90": 0.421097971385433, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652752 }, "cmh_m_s": { "median": 0.03739314717898144, "p90": 0.059713013430113884, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652752 }, "rhcm_continuous_estimate_m_s": { "median": 0.1677584054809312, "p90": 0.40292928792015537, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652752 }, "rhcm_discrete_m_s": { "median": 0.1678806694101696, "p90": 0.402868649936001, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652752 }, "wms": { "median": 0.8012456362588181, "p90": 0.9520758104339802, "valid_pairs": 489, "valid_duration_s": 16.299999999999997, "coverage": 0.8163606010016692 }, "rham_deg_s": { "median": 33.59917966823078, "p90": 85.75899456530937, "valid_pairs": 491, "valid_duration_s": 16.366666666666664, "coverage": 0.8196994991652752 }, "speed_spearman": 0.15076078218282493, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1960128492764117, "p90": 0.45460858493858963, "valid_pairs": 517, "valid_duration_s": 17.233333333333334, "coverage": 0.8631051752921536 }, "cmh_m_s": { "median": 0.03816622026997649, "p90": 0.08373826046381078, "valid_pairs": 517, "valid_duration_s": 17.233333333333334, "coverage": 0.8631051752921536 }, "rhcm_continuous_estimate_m_s": { "median": 0.20024172709942806, "p90": 0.42952212244784505, "valid_pairs": 517, "valid_duration_s": 17.233333333333334, "coverage": 0.8631051752921536 }, "rhcm_discrete_m_s": { "median": 0.20028308627057675, "p90": 0.4294720789775846, "valid_pairs": 517, "valid_duration_s": 17.233333333333334, "coverage": 0.8631051752921536 }, "wms": { "median": 0.7862294307385658, "p90": 0.966948684453118, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285474 }, "rham_deg_s": { "median": 33.360665761328164, "p90": 80.9215007990017, "valid_pairs": 517, "valid_duration_s": 17.233333333333334, "coverage": 0.8631051752921536 }, "speed_spearman": 0.32156998757691485, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.057376760842696134, "camera_angular_speed_p90_deg_s": 14.975154036443177, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.45460858493858963, "hand_camera_speed_p90_m_s": 0.4294720789775846, "hand_world_valid_fraction": 0.85, "hand_camera_valid_fraction": 0.85, "cmh_p90_m_s": 0.08373826046381078, "rhcm_p90_m_s": 0.4294720789775846, "rham_p90_deg_s": 85.75899456530937, "wms_p90": 0.966948684453118, "gt_coverage": 0.8631051752921536, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "231b9f6dbdf6f22315a5e93f3afce5a965e481bbd5f272e134851387d757c232", "metadata_sha256": "ebe60b3c637093b8c78003c69ce406bb976ad5d9a21d84b2b0871f9651cde8b7", "source_file": "long20_20260918/local/train/previews/46d68e78e2b44ae5a011094e/reference.npz", "window_s": [ 2.2, 22.166666666666668 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.552252871008668e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "605a5052a0a3d7881cac88c5df22e5d53bae57df7254afe7beda23f850971824", "mano_sha256": "0e4c04f4b597fd5045f8afa87f917c6b2185b91fd7cc0510898ecfed7fcd15c7", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "4c46b2108f000cd6f5f0d236", "sample_id": "ef864e911da89db1bbf464b8", "file": "20_seconds/ARCTIC/ef864e911da89db1bbf464b8_MANO_GT_20.00s.mp4", "recording_id": "s06/scissors_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 13/600 output frames; may include genuine out-of-view frames", "id": "4c46b2108f000cd6f5f0d236", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/ef864e911da89db1bbf464b8_MANO_GT_20.00s.mp4", "bytes": 4400946, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.027001209083949287, "p90": 0.08193504316400722, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 4.324785448447642, "p90": 14.190464531913017, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.05428034081393898, "p90": 0.20777582199542327, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "cmh_m_s": { "median": 0.03516570951083306, "p90": 0.12184952060900592, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "rhcm_continuous_estimate_m_s": { "median": 0.07404347765840857, "p90": 0.23258700862542114, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "rhcm_discrete_m_s": { "median": 0.07403999466229388, "p90": 0.23258442541591492, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "wms": { "median": 0.7668354551656111, "p90": 0.9625252930058101, "valid_pairs": 526, "valid_duration_s": 17.533333333333335, "coverage": 0.8781302170283808 }, "rham_deg_s": { "median": 46.46956960797383, "p90": 171.0063503213674, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "speed_spearman": 0.432370093650552, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.038547335407110576, "p90": 0.14253128623051128, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.9732888146911521 }, "cmh_m_s": { "median": 0.03670193984856739, "p90": 0.11564030658741355, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.9732888146911521 }, "rhcm_continuous_estimate_m_s": { "median": 0.04857817597983146, "p90": 0.16796227295169694, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.9732888146911521 }, "rhcm_discrete_m_s": { "median": 0.04860420640396928, "p90": 0.16832077604881102, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.9732888146911521 }, "wms": { "median": 0.6713844575880428, "p90": 0.908573463947285, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549249 }, "rham_deg_s": { "median": 26.143332436773907, "p90": 140.42095561291893, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.9732888146911521 }, "speed_spearman": 0.4809944254464935, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.08193504316400722, "camera_angular_speed_p90_deg_s": 14.190464531913017, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.20777582199542327, "hand_camera_speed_p90_m_s": 0.23258442541591492, "hand_world_valid_fraction": 0.9283333333333333, "hand_camera_valid_fraction": 0.9283333333333333, "cmh_p90_m_s": 0.12184952060900592, "rhcm_p90_m_s": 0.23258442541591492, "rham_p90_deg_s": 171.0063503213674, "wms_p90": 0.9625252930058101, "gt_coverage": 0.9732888146911521, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "768655e713f8b6a41d2b21cb8e524940275ad6965f3ad0413e34fd0208dc6bed", "metadata_sha256": "10cda4ed82f00bbc837e8e8a5db0e1a263f58d8eb6830a2c04d677f855365f0f", "source_file": "long20_20260918/local/train/previews/4c46b2108f000cd6f5f0d236/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.526797389797801e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "94214d7cbb23aaa78914269d5e06792bbdaa27e9ad4ed602d03ede1b8eae21dd", "mano_sha256": "967ec186dd8a832983e8342e6163eb6e5602d91d7578e977a0e4e1df8f461e08", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "4c850533cc98eb45d70d07d0", "sample_id": "31eafe2ee8f6a190fb8427eb", "file": "20_seconds/ARCTIC/31eafe2ee8f6a190fb8427eb_MANO_GT_20.00s.mp4", "recording_id": "s09/scissors_use_04", "source_first_frame": "52", "source_last_frame": "651", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 59/600 output frames; may include genuine out-of-view frames", "id": "4c850533cc98eb45d70d07d0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/31eafe2ee8f6a190fb8427eb_MANO_GT_20.00s.mp4", "bytes": 3744738, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.014076753025298708, "p90": 0.03221205365845915, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 2.6953895749290444, "p90": 5.627709183012627, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.024573626162445197, "p90": 0.11012866359117168, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "cmh_m_s": { "median": 0.021869833875060588, "p90": 0.045627098655008996, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "rhcm_continuous_estimate_m_s": { "median": 0.032196433914299195, "p90": 0.11695133235761536, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "rhcm_discrete_m_s": { "median": 0.03216515689534031, "p90": 0.11695390924863779, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "wms": { "median": 0.7153106565126488, "p90": 0.9243709820135895, "valid_pairs": 504, "valid_duration_s": 16.799999999999997, "coverage": 0.8414023372287145 }, "rham_deg_s": { "median": 22.579097010358645, "p90": 110.20521474207386, "valid_pairs": 535, "valid_duration_s": 17.833333333333336, "coverage": 0.8931552587646079 }, "speed_spearman": 0.30525397270708304, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.03214397468922412, "p90": 0.13804912334577887, "valid_pairs": 540, "valid_duration_s": 18.0, "coverage": 0.9015025041736228 }, "cmh_m_s": { "median": 0.022709561587948762, "p90": 0.04757679097783425, "valid_pairs": 540, "valid_duration_s": 18.0, "coverage": 0.9015025041736228 }, "rhcm_continuous_estimate_m_s": { "median": 0.04159973230086762, "p90": 0.144434324582266, "valid_pairs": 540, "valid_duration_s": 18.0, "coverage": 0.9015025041736228 }, "rhcm_discrete_m_s": { "median": 0.041586214557650165, "p90": 0.14443270378919446, "valid_pairs": 540, "valid_duration_s": 18.0, "coverage": 0.9015025041736228 }, "wms": { "median": 0.7318674395954007, "p90": 0.9530184329095909, "valid_pairs": 520, "valid_duration_s": 17.333333333333336, "coverage": 0.8681135225375628 }, "rham_deg_s": { "median": 21.667380614478727, "p90": 61.687049108885496, "valid_pairs": 540, "valid_duration_s": 18.0, "coverage": 0.9015025041736228 }, "speed_spearman": 0.2576728848711949, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.03221205365845915, "camera_angular_speed_p90_deg_s": 5.627709183012627, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.13804912334577887, "hand_camera_speed_p90_m_s": 0.14443270378919446, "hand_world_valid_fraction": 0.8975, "hand_camera_valid_fraction": 0.8975, "cmh_p90_m_s": 0.04757679097783425, "rhcm_p90_m_s": 0.14443270378919446, "rham_p90_deg_s": 110.20521474207386, "wms_p90": 0.9530184329095909, "gt_coverage": 0.9015025041736228, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "0c4a49cb213e0873a42ddb10c7f535934ac0903a5cf877ca4476a38f6216b476", "metadata_sha256": "4ce9c720db04d780643c171648eb244a625f5795699e420db43e6add48c501b1", "source_file": "long20_20260918/local/train/previews/4c850533cc98eb45d70d07d0/reference.npz", "window_s": [ 1.6666666666666667, 21.633333333333333 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.585973723508303e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "dc7726a4cdd5c565b89dabaf415c280dc41c522f0e9b01775527eae1bff39633", "mano_sha256": "87843a239c60fe5b5dc9bbcf08a8491631ae3d52891870b8f4f73f96a43de8db", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "4dc6fc0d05022dfd9713f695", "sample_id": "96622be6fe508fdc2d95467f", "file": "20_seconds/ARCTIC/96622be6fe508fdc2d95467f_MANO_GT_20.00s.mp4", "recording_id": "s07/espressomachine_use_01", "source_first_frame": "2", "source_last_frame": "601", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 16/600 output frames; may include genuine out-of-view frames", "id": "4dc6fc0d05022dfd9713f695", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/96622be6fe508fdc2d95467f_MANO_GT_20.00s.mp4", "bytes": 5133690, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03950663764004062, "p90": 0.09259440043056903, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 6.790057948021076, "p90": 17.27327961039177, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.059815991946668486, "p90": 0.2510134816935634, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "cmh_m_s": { "median": 0.034078310092131116, "p90": 0.09076515976828287, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "rhcm_continuous_estimate_m_s": { "median": 0.07529713951091917, "p90": 0.24965058008616833, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "rhcm_discrete_m_s": { "median": 0.07507811066548799, "p90": 0.2497519541042653, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "wms": { "median": 0.7793612138081658, "p90": 0.9532793601283246, "valid_pairs": 504, "valid_duration_s": 16.800000000000004, "coverage": 0.8414023372287148 }, "rham_deg_s": { "median": 28.76207053505672, "p90": 78.55071757316071, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "speed_spearman": 0.09308095059813387, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.096656595361608, "p90": 0.2458060877808646, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "cmh_m_s": { "median": 0.030898411582010087, "p90": 0.06095557417854433, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rhcm_continuous_estimate_m_s": { "median": 0.0989509187308868, "p90": 0.24096983905014607, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rhcm_discrete_m_s": { "median": 0.09896484083630666, "p90": 0.24097253305236127, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "wms": { "median": 0.7895268177804007, "p90": 0.9565033051446916, "valid_pairs": 575, "valid_duration_s": 19.166666666666664, "coverage": 0.9599332220367278 }, "rham_deg_s": { "median": 30.223560860380648, "p90": 81.03410041098473, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "speed_spearman": 0.16439196995956096, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.09259440043056903, "camera_angular_speed_p90_deg_s": 17.27327961039177, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2510134816935634, "hand_camera_speed_p90_m_s": 0.2497519541042653, "hand_world_valid_fraction": 0.9091666666666667, "hand_camera_valid_fraction": 0.9091666666666667, "cmh_p90_m_s": 0.09076515976828287, "rhcm_p90_m_s": 0.2497519541042653, "rham_p90_deg_s": 81.03410041098473, "wms_p90": 0.9565033051446916, "gt_coverage": 0.963272120200334, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "87771ab9ced3e15bd3f2607e6555fc14562190bbff3969fd0927117766132530", "metadata_sha256": "bc529589e010c6b7f919ac5e6ce8e7d958f21d7df5f220f2ebbd39840dd3751e", "source_file": "long20_20260918/local/train/previews/4dc6fc0d05022dfd9713f695/reference.npz", "window_s": [ 0.03333333333333333, 20.0 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.585973723508303e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "1a6d7a7b14e02ee839b823cc9c93537c809579fc80acb03c0d3cf03d3f8beab7", "mano_sha256": "26b74b32384446da21f4345137d6fe7d6e77f0520fe9b640eef5e026290849eb", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "4e8a067f2e5c5a59248dd1fa", "sample_id": "6ac6cb71ea77eb79654b0116", "file": "20_seconds/ARCTIC/6ac6cb71ea77eb79654b0116_MANO_GT_20.00s.mp4", "recording_id": "s01/notebook_use_02", "source_first_frame": "1", "source_last_frame": "600", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 31/600 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 7 frames", "id": "4e8a067f2e5c5a59248dd1fa", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/6ac6cb71ea77eb79654b0116_MANO_GT_20.00s.mp4", "bytes": 6591899, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.032601732951131425, "p90": 0.07422226211006118, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 6.6376360779415355, "p90": 16.057351368622733, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.1365861780920172, "p90": 0.3758118617786189, "valid_pairs": 565, "valid_duration_s": 18.833333333333332, "coverage": 0.9432387312186978 }, "cmh_m_s": { "median": 0.04183058920622543, "p90": 0.09733213863760379, "valid_pairs": 565, "valid_duration_s": 18.833333333333332, "coverage": 0.9432387312186978 }, "rhcm_continuous_estimate_m_s": { "median": 0.15016457158703458, "p90": 0.4106209088913115, "valid_pairs": 565, "valid_duration_s": 18.833333333333332, "coverage": 0.9432387312186978 }, "rhcm_discrete_m_s": { "median": 0.1502394686614731, "p90": 0.4122956681995149, "valid_pairs": 565, "valid_duration_s": 18.833333333333332, "coverage": 0.9432387312186978 }, "wms": { "median": 0.8330247724069783, "p90": 0.9784353398452882, "valid_pairs": 565, "valid_duration_s": 18.833333333333332, "coverage": 0.9432387312186978 }, "rham_deg_s": { "median": 83.90253863528332, "p90": 248.94048603235188, "valid_pairs": 565, "valid_duration_s": 18.833333333333332, "coverage": 0.9432387312186978 }, "speed_spearman": 0.15024693731857278, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.10996393722227667, "p90": 0.3124804663416792, "valid_pairs": 565, "valid_duration_s": 18.83333333333333, "coverage": 0.9432387312186977 }, "cmh_m_s": { "median": 0.03896084413405856, "p90": 0.09910573037429277, "valid_pairs": 565, "valid_duration_s": 18.83333333333333, "coverage": 0.9432387312186977 }, "rhcm_continuous_estimate_m_s": { "median": 0.1220719640900952, "p90": 0.3321871154777421, "valid_pairs": 565, "valid_duration_s": 18.83333333333333, "coverage": 0.9432387312186977 }, "rhcm_discrete_m_s": { "median": 0.12205587202092369, "p90": 0.33218676499321326, "valid_pairs": 565, "valid_duration_s": 18.83333333333333, "coverage": 0.9432387312186977 }, "wms": { "median": 0.8126819467912809, "p90": 0.9584351611637957, "valid_pairs": 564, "valid_duration_s": 18.799999999999997, "coverage": 0.9415692821368947 }, "rham_deg_s": { "median": 74.2884936919704, "p90": 213.62084875589244, "valid_pairs": 565, "valid_duration_s": 18.83333333333333, "coverage": 0.9432387312186977 }, "speed_spearman": 0.24107429543190914, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.07422226211006118, "camera_angular_speed_p90_deg_s": 16.057351368622733, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3758118617786189, "hand_camera_speed_p90_m_s": 0.4122956681995149, "hand_world_valid_fraction": 0.9466666666666667, "hand_camera_valid_fraction": 0.9466666666666667, "cmh_p90_m_s": 0.09910573037429277, "rhcm_p90_m_s": 0.4122956681995149, "rham_p90_deg_s": 248.94048603235188, "wms_p90": 0.9784353398452882, "gt_coverage": 0.9432387312186978, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "d38e383cd3f05e9438fdb846fa88546dc72aa9abec24c47f31d5c14d10acd2b1", "metadata_sha256": "54726c26e1930960e913ea3cd200be1aec87aad13edac6c783967cba04fa42d7", "source_file": "long20_20260918/local/train/previews/4e8a067f2e5c5a59248dd1fa/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.52041112178401e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "831e5856fab8c2634dea28433ad90e02b7265f36a3f725eb78029490ef7a5e39", "mano_sha256": "60524279d9827fbc050a8094563c72f097bf56a49d4fa245e72163c5aa652bb2", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "50fd99a096da624deeff3d05", "sample_id": "0856ec44b1451424aeef61ba", "file": "20_seconds/ARCTIC/0856ec44b1451424aeef61ba_MANO_GT_20.00s.mp4", "recording_id": "s02/phone_grab_01", "source_first_frame": "51", "source_last_frame": "650", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 59/600 output frames; may include genuine out-of-view frames", "id": "50fd99a096da624deeff3d05", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/0856ec44b1451424aeef61ba_MANO_GT_20.00s.mp4", "bytes": 4230120, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.026016851859117465, "p90": 0.04937442738225013, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 5.684421935179569, "p90": 11.938702615818979, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.12283043398570505, "p90": 0.4791386466971781, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "cmh_m_s": { "median": 0.038167049785323974, "p90": 0.0700540171794549, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "rhcm_continuous_estimate_m_s": { "median": 0.11508748420856074, "p90": 0.46807960495418927, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "rhcm_discrete_m_s": { "median": 0.11506524656764947, "p90": 0.46813164333424157, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "wms": { "median": 0.7314367317933808, "p90": 0.9488482583584985, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "rham_deg_s": { "median": 66.2685222760454, "p90": 196.91698026845881, "valid_pairs": 539, "valid_duration_s": 17.96666666666667, "coverage": 0.8998330550918197 }, "speed_spearman": 0.06150312778034498, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04937442738225013, "camera_angular_speed_p90_deg_s": 11.938702615818979, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4791386466971781, "hand_camera_speed_p90_m_s": 0.46813164333424157, "hand_world_valid_fraction": 0.4508333333333333, "hand_camera_valid_fraction": 0.4508333333333333, "cmh_p90_m_s": 0.0700540171794549, "rhcm_p90_m_s": 0.46813164333424157, "rham_p90_deg_s": 196.91698026845881, "wms_p90": 0.9488482583584985, "gt_coverage": 0.8998330550918197, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "54b5ed3c2f1c6e85758c2cf2e3c3719eca2f46b9cb16577f093c39e82cd68030", "metadata_sha256": "02aeb4618823a679d652e600a2eb7928d31ce2d10f3e8bd36de5c4f030c7f8aa", "source_file": "long20_20260918/local/train/previews/50fd99a096da624deeff3d05/reference.npz", "window_s": [ 1.6333333333333333, 21.6 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.449557262054371e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b679cc933ed83e22e06bc84404ca322be77cb1654cd052c790fe3dac3af2e2d1", "mano_sha256": "b68f09c9f0874ba8fa9430a08c4bb53de7420c77977039cfbcbe1ac88aa15361", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "False", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "5702df3fc172409778607b90", "sample_id": "e1ee035e2da75fc2c200aaa0", "file": "20_seconds/ARCTIC/e1ee035e2da75fc2c200aaa0_MANO_GT_20.00s.mp4", "recording_id": "s05/mixer_use_01", "source_first_frame": "64", "source_last_frame": "663", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 55/600 output frames; may include genuine out-of-view frames", "id": "5702df3fc172409778607b90", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/e1ee035e2da75fc2c200aaa0_MANO_GT_20.00s.mp4", "bytes": 5393180, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.06791433630840607, "p90": 0.20599828750106242, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 13.517323385215718, "p90": 26.183350777851306, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.13233970026203373, "p90": 0.31222559006087425, "valid_pairs": 523, "valid_duration_s": 17.433333333333334, "coverage": 0.8731218697829717 }, "cmh_m_s": { "median": 0.08313322325694224, "p90": 0.16683993688263626, "valid_pairs": 523, "valid_duration_s": 17.433333333333334, "coverage": 0.8731218697829717 }, "rhcm_continuous_estimate_m_s": { "median": 0.15821968503180517, "p90": 0.31042808730564897, "valid_pairs": 523, "valid_duration_s": 17.433333333333334, "coverage": 0.8731218697829717 }, "rhcm_discrete_m_s": { "median": 0.1585029645512691, "p90": 0.30988017375666205, "valid_pairs": 523, "valid_duration_s": 17.433333333333334, "coverage": 0.8731218697829717 }, "wms": { "median": 0.7631471552495911, "p90": 0.9609327606778227, "valid_pairs": 523, "valid_duration_s": 17.433333333333334, "coverage": 0.8731218697829717 }, "rham_deg_s": { "median": 54.19704175454069, "p90": 135.28878348503127, "valid_pairs": 523, "valid_duration_s": 17.433333333333334, "coverage": 0.8731218697829717 }, "speed_spearman": -0.01721185223510732, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.12712192815741125, "p90": 0.30380461176043383, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "cmh_m_s": { "median": 0.08479606095250035, "p90": 0.17919578376742623, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "rhcm_continuous_estimate_m_s": { "median": 0.14586294634842298, "p90": 0.3541466850143872, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "rhcm_discrete_m_s": { "median": 0.14607036792581451, "p90": 0.3543926004663795, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "wms": { "median": 0.6809655318690816, "p90": 0.951009477111397, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "rham_deg_s": { "median": 62.35752147244843, "p90": 155.2613029975367, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "speed_spearman": 0.2911346921385721, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.20599828750106242, "camera_angular_speed_p90_deg_s": 26.183350777851306, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.31222559006087425, "hand_camera_speed_p90_m_s": 0.3543926004663795, "hand_world_valid_fraction": 0.8841666666666667, "hand_camera_valid_fraction": 0.8841666666666667, "cmh_p90_m_s": 0.17919578376742623, "rhcm_p90_m_s": 0.3543926004663795, "rham_p90_deg_s": 155.2613029975367, "wms_p90": 0.9609327606778227, "gt_coverage": 0.8898163606010017, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "c5d57ad84c21c82102d90c3517ed867c5d5b1293002ab5d137af7915b78e8e41", "metadata_sha256": "3b64895a677595e3995b22406cf11093adc13948864a85a8963bbf3a4fbe6414", "source_file": "long20_20260918/local/benchmark/previews/5702df3fc172409778607b90/reference.npz", "window_s": [ 2.1, 22.066666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.660953740672398e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "bb8bf44503453dce44a48f3187c72e195f0d51e2f64c437164637a4244849658", "mano_sha256": "3c5a323882a0c07c1e95f8533e2f7ed279db14f7ecc1007a2d8c36eb1d61ad24", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "579127f9c4e0496b17db02b6", "sample_id": "fac834e35f2d141335f8dde4", "file": "20_seconds/ARCTIC/fac834e35f2d141335f8dde4_MANO_GT_20.00s.mp4", "recording_id": "s10/notebook_use_02", "source_first_frame": "105", "source_last_frame": "704", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 69/600 output frames; may include genuine out-of-view frames", "id": "579127f9c4e0496b17db02b6", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/fac834e35f2d141335f8dde4_MANO_GT_20.00s.mp4", "bytes": 4663474, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.04360554403060534, "p90": 0.09384793304065767, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 6.0126136670017685, "p90": 12.882541936559608, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08696139578640788, "p90": 0.18728767877701283, "valid_pairs": 181, "valid_duration_s": 6.033333333333333, "coverage": 0.30217028380634386 }, "cmh_m_s": { "median": 0.05085618562901657, "p90": 0.09936505119844538, "valid_pairs": 181, "valid_duration_s": 6.033333333333333, "coverage": 0.30217028380634386 }, "rhcm_continuous_estimate_m_s": { "median": 0.08836422468940605, "p90": 0.1988515084510493, "valid_pairs": 181, "valid_duration_s": 6.033333333333333, "coverage": 0.30217028380634386 }, "rhcm_discrete_m_s": { "median": 0.088371440573314, "p90": 0.19902856331161167, "valid_pairs": 181, "valid_duration_s": 6.033333333333333, "coverage": 0.30217028380634386 }, "wms": { "median": 0.6778860261604491, "p90": 0.9206035910850285, "valid_pairs": 180, "valid_duration_s": 6.0, "coverage": 0.30050083472454087 }, "rham_deg_s": { "median": 45.55717311965335, "p90": 132.78889556418184, "valid_pairs": 181, "valid_duration_s": 6.033333333333333, "coverage": 0.30217028380634386 }, "speed_spearman": 0.25000910691518424, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.059365461963839315, "p90": 0.30188030757869877, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113521 }, "cmh_m_s": { "median": 0.04027485086973315, "p90": 0.09277915943122655, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113521 }, "rhcm_continuous_estimate_m_s": { "median": 0.06726992393956421, "p90": 0.3012605758346551, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113521 }, "rhcm_discrete_m_s": { "median": 0.0671958306391969, "p90": 0.30129488597107024, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113521 }, "wms": { "median": 0.7066797369601538, "p90": 0.9290870976889797, "valid_pairs": 433, "valid_duration_s": 14.433333333333337, "coverage": 0.7228714524207013 }, "rham_deg_s": { "median": 29.998762217012427, "p90": 150.09398644119702, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113521 }, "speed_spearman": 0.5164381058671895, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.09384793304065767, "camera_angular_speed_p90_deg_s": 12.882541936559608, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.30188030757869877, "hand_camera_speed_p90_m_s": 0.30129488597107024, "hand_world_valid_fraction": 0.535, "hand_camera_valid_fraction": 0.535, "cmh_p90_m_s": 0.09936505119844538, "rhcm_p90_m_s": 0.30129488597107024, "rham_p90_deg_s": 150.09398644119702, "wms_p90": 0.9290870976889797, "gt_coverage": 0.7512520868113521, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "caad42dbe1b098a44513b5faccde77ecffde7d4035ad9003177015e5903f1fa6", "metadata_sha256": "b818dbf8ace86ee708f216bbc4ef2937bc573abf66751e369859047a9f8b4fe0", "source_file": "long20_20260918/local/train/previews/579127f9c4e0496b17db02b6/reference.npz", "window_s": [ 3.466666666666667, 23.433333333333334 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.577566798522237e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6b9ec42e2f6836a4d39dfddf1b836511fdd5a615e4693d9d964c83f18f1a9fea", "mano_sha256": "6e05f1cf4a434d9b3e3aa305a6135266b237ab068f35f4f4f08aba07c60cd85e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "5d2302f8779fde7e11a4bfe8", "sample_id": "af26765d803ea2a5494056a1", "file": "20_seconds/ARCTIC/af26765d803ea2a5494056a1_MANO_GT_20.00s.mp4", "recording_id": "s05/notebook_use_01", "source_first_frame": "76", "source_last_frame": "675", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 5/600 output frames; may include genuine out-of-view frames", "id": "5d2302f8779fde7e11a4bfe8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/af26765d803ea2a5494056a1_MANO_GT_20.00s.mp4", "bytes": 6456375, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.07682054143043178, "p90": 0.16555323422493506, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 18.65836217940564, "p90": 40.983479682661624, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.11078390539919283, "p90": 0.25784789730452534, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5308848080133556 }, "cmh_m_s": { "median": 0.09258995713029418, "p90": 0.2215722289768363, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5308848080133556 }, "rhcm_continuous_estimate_m_s": { "median": 0.15804722118851505, "p90": 0.30166409106928077, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5308848080133556 }, "rhcm_discrete_m_s": { "median": 0.15822242322531777, "p90": 0.30191392520507837, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5308848080133556 }, "wms": { "median": 0.701574808885315, "p90": 0.940824151417492, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5308848080133556 }, "rham_deg_s": { "median": 81.212374624863, "p90": 262.1148494765628, "valid_pairs": 318, "valid_duration_s": 10.6, "coverage": 0.5308848080133556 }, "speed_spearman": 0.11838095918682255, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.07969153589904471, "p90": 0.1915646281365455, "valid_pairs": 552, "valid_duration_s": 18.4, "coverage": 0.9215358931552587 }, "cmh_m_s": { "median": 0.09013442299388062, "p90": 0.1912373866765952, "valid_pairs": 552, "valid_duration_s": 18.4, "coverage": 0.9215358931552587 }, "rhcm_continuous_estimate_m_s": { "median": 0.13987170712644748, "p90": 0.2898568415674307, "valid_pairs": 552, "valid_duration_s": 18.4, "coverage": 0.9215358931552587 }, "rhcm_discrete_m_s": { "median": 0.1395576438108857, "p90": 0.28882967122670855, "valid_pairs": 552, "valid_duration_s": 18.4, "coverage": 0.9215358931552587 }, "wms": { "median": 0.7636854636057199, "p90": 0.937129223944076, "valid_pairs": 552, "valid_duration_s": 18.4, "coverage": 0.9215358931552587 }, "rham_deg_s": { "median": 57.861919542580466, "p90": 145.5249766415719, "valid_pairs": 552, "valid_duration_s": 18.4, "coverage": 0.9215358931552587 }, "speed_spearman": 0.13672629755361207, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.16555323422493506, "camera_angular_speed_p90_deg_s": 40.983479682661624, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.25784789730452534, "hand_camera_speed_p90_m_s": 0.30191392520507837, "hand_world_valid_fraction": 0.7375, "hand_camera_valid_fraction": 0.7375, "cmh_p90_m_s": 0.2215722289768363, "rhcm_p90_m_s": 0.30191392520507837, "rham_p90_deg_s": 262.1148494765628, "wms_p90": 0.940824151417492, "gt_coverage": 0.9215358931552587, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "1a6d1fd2300e8512de2567e7d140fc015e9de00cf033c26703c0c2f191673200", "metadata_sha256": "1d595d0fd759fd1e3409684df1f7a8a603ef528a25c3f2d5cb34b1c997b7302b", "source_file": "long20_20260918/local/benchmark/previews/5d2302f8779fde7e11a4bfe8/reference.npz", "window_s": [ 2.5, 22.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.579669976578771e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "da3914cc02c0deb5ddc1a162c9fc65d7280018282802fb540c1588ea832d8170", "mano_sha256": "79e50a62106c83a7b9bc0e7a94aa87014ecd7c01128c3cbf7e21c648986092f8", "duration_s": 19.5, "frames": 585, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "6495e6b445ee0eadb915a3f0", "sample_id": "10c036e2933da33287de4a63", "file": "short_originals_under_20s/ARCTIC/10c036e2933da33287de4a63_MANO_GT_19.50s.mp4", "recording_id": "s09/microwave_use_03", "source_first_frame": "2", "source_last_frame": "586", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 65/585 output frames; may include genuine out-of-view frames", "id": "6495e6b445ee0eadb915a3f0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/ARCTIC/10c036e2933da33287de4a63_MANO_GT_19.50s.mp4", "bytes": 3794203, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.018770631622869745, "p90": 0.05010405055855692, "valid_pairs": 584, "valid_duration_s": 19.466666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 4.43585392523898, "p90": 8.760365233021172, "valid_pairs": 584, "valid_duration_s": 19.466666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.09736618209567362, "p90": 0.4284994572630882, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.875 }, "cmh_m_s": { "median": 0.04001280456460471, "p90": 0.08650239025773351, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.875 }, "rhcm_continuous_estimate_m_s": { "median": 0.09423191094241307, "p90": 0.4015004349137308, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.875 }, "rhcm_discrete_m_s": { "median": 0.09432137052391679, "p90": 0.40175316758971735, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.875 }, "wms": { "median": 0.7469803959577768, "p90": 0.9485158803496738, "valid_pairs": 505, "valid_duration_s": 16.833333333333336, "coverage": 0.8647260273972605 }, "rham_deg_s": { "median": 40.15116318953833, "p90": 168.05786998622096, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.875 }, "speed_spearman": 0.39442198112121557, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.00864083983810702, "p90": 0.028836528408441024, "valid_pairs": 519, "valid_duration_s": 17.299999999999997, "coverage": 0.8886986301369862 }, "cmh_m_s": { "median": 0.04008949411776174, "p90": 0.09009598302232415, "valid_pairs": 519, "valid_duration_s": 17.299999999999997, "coverage": 0.8886986301369862 }, "rhcm_continuous_estimate_m_s": { "median": 0.04212400640678309, "p90": 0.10363603893735122, "valid_pairs": 519, "valid_duration_s": 17.299999999999997, "coverage": 0.8886986301369862 }, "rhcm_discrete_m_s": { "median": 0.04212858235077268, "p90": 0.10357918795161854, "valid_pairs": 519, "valid_duration_s": 17.299999999999997, "coverage": 0.8886986301369862 }, "wms": { "median": 0.8488228324477841, "p90": 0.9588082937853009, "valid_pairs": 487, "valid_duration_s": 16.233333333333334, "coverage": 0.8339041095890413 }, "rham_deg_s": { "median": 8.556339480804494, "p90": 28.076507147815242, "valid_pairs": 519, "valid_duration_s": 17.299999999999997, "coverage": 0.8886986301369862 }, "speed_spearman": 0.2817246961755632, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05010405055855692, "camera_angular_speed_p90_deg_s": 8.760365233021172, "camera_valid_fraction": 1.0, "camera_pairs": 584, "hand_world_speed_p90_m_s": 0.4284994572630882, "hand_camera_speed_p90_m_s": 0.40175316758971735, "hand_world_valid_fraction": 0.882051282051282, "hand_camera_valid_fraction": 0.882051282051282, "cmh_p90_m_s": 0.09009598302232415, "rhcm_p90_m_s": 0.40175316758971735, "rham_p90_deg_s": 168.05786998622096, "wms_p90": 0.9588082937853009, "gt_coverage": 0.8886986301369862, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "d135b59e3e119913f7acc7868700e7b7e201c4b779718c5feedc913ae46ea9aa", "metadata_sha256": "e06ee59b7a23fd23163534e2a18ac2ec186a2348bceadb9e04e8f54b1a1fd4a5", "source_file": "long20_20260918/local/train/previews/6495e6b445ee0eadb915a3f0/reference.npz", "window_s": [ 0.0, 19.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 585, "common_world_residual_max_m": 4.75098925995612e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b0e4a370dff78167da849ea028fbe9e638248e1ffde55113fcf37fd6b2ac7202", "mano_sha256": "779135e159209b626f4ea31cb905d7d775ca739ea70eef8b96f52ac92302c2e8", "duration_s": 19.834, "frames": 595, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "662d35d433c231b957eaceef", "sample_id": "e7ae2de6f7b2f7feea2cf463", "file": "short_originals_under_20s/ARCTIC/e7ae2de6f7b2f7feea2cf463_MANO_GT_19.83s.mp4", "recording_id": "s09/espressomachine_use_04", "source_first_frame": "2", "source_last_frame": "596", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 71/595 output frames; may include genuine out-of-view frames", "id": "662d35d433c231b957eaceef", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/ARCTIC/e7ae2de6f7b2f7feea2cf463_MANO_GT_19.83s.mp4", "bytes": 3625079, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.017822324858868872, "p90": 0.05267430753941304, "valid_pairs": 594, "valid_duration_s": 19.8, "coverage": 1.0 }, "camera_rotation": { "median": 3.5240075704380427, "p90": 8.131670357116151, "valid_pairs": 594, "valid_duration_s": 19.8, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0056978909229309545, "p90": 0.015715630406019393, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8804713804713806 }, "cmh_m_s": { "median": 0.025233611808270818, "p90": 0.053610867744348575, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8804713804713806 }, "rhcm_continuous_estimate_m_s": { "median": 0.025227079288887306, "p90": 0.0504249706606574, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8804713804713806 }, "rhcm_discrete_m_s": { "median": 0.025204593911373422, "p90": 0.050440987526436255, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8804713804713806 }, "wms": { "median": 0.8158490263485567, "p90": 0.9357046884466551, "valid_pairs": 445, "valid_duration_s": 14.83333333333334, "coverage": 0.7491582491582495 }, "rham_deg_s": { "median": 5.352110484498098, "p90": 16.15831902959986, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8804713804713806 }, "speed_spearman": 0.22336821386863237, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.16923797935917814, "p90": 0.3264623633213865, "valid_pairs": 519, "valid_duration_s": 17.3, "coverage": 0.8737373737373737 }, "cmh_m_s": { "median": 0.027463283708671266, "p90": 0.054602240160857705, "valid_pairs": 519, "valid_duration_s": 17.3, "coverage": 0.8737373737373737 }, "rhcm_continuous_estimate_m_s": { "median": 0.17014851357194727, "p90": 0.3360013306656355, "valid_pairs": 519, "valid_duration_s": 17.3, "coverage": 0.8737373737373737 }, "rhcm_discrete_m_s": { "median": 0.1701496187288483, "p90": 0.33595121181669174, "valid_pairs": 519, "valid_duration_s": 17.3, "coverage": 0.8737373737373737 }, "wms": { "median": 0.874731777699729, "p90": 0.9800316792106831, "valid_pairs": 518, "valid_duration_s": 17.266666666666666, "coverage": 0.872053872053872 }, "rham_deg_s": { "median": 36.44171514444441, "p90": 119.25862436960963, "valid_pairs": 519, "valid_duration_s": 17.3, "coverage": 0.8737373737373737 }, "speed_spearman": 0.18173602682273207, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05267430753941304, "camera_angular_speed_p90_deg_s": 8.131670357116151, "camera_valid_fraction": 1.0, "camera_pairs": 594, "hand_world_speed_p90_m_s": 0.3264623633213865, "hand_camera_speed_p90_m_s": 0.33595121181669174, "hand_world_valid_fraction": 0.8773109243697479, "hand_camera_valid_fraction": 0.8773109243697479, "cmh_p90_m_s": 0.054602240160857705, "rhcm_p90_m_s": 0.33595121181669174, "rham_p90_deg_s": 119.25862436960963, "wms_p90": 0.9800316792106831, "gt_coverage": 0.8804713804713806, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "459be53f425b8cc013cfc1b1d2addebe7c5ce1a396cfb3b0479f99fb3ddb4cea", "metadata_sha256": "860acc997335caadd5c25c1bb53210a2e032ac0ab4ba1d1f9d9f38141a27b4b6", "source_file": "long20_20260918/local/train/previews/662d35d433c231b957eaceef/reference.npz", "window_s": [ 0.0, 19.8 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 595, "common_world_residual_max_m": 4.710277376051325e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ae5bca4d51bc4ca65d0a747df41e20fafbfef3fe008ebd47a1a490c5d9c5d96b", "mano_sha256": "cee736921c57b29ae06e7b54ee4f4aad1ae00abf7120df1f801dc2428f8ba815", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "6644e508bb0fdf63e6777f28", "sample_id": "d866cbacd22439b6159cfdfa", "file": "20_seconds/ARCTIC/d866cbacd22439b6159cfdfa_MANO_GT_20.00s.mp4", "recording_id": "s05/capsulemachine_use_02", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 25/600 output frames; may include genuine out-of-view frames", "id": "6644e508bb0fdf63e6777f28", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/d866cbacd22439b6159cfdfa_MANO_GT_20.00s.mp4", "bytes": 5494749, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.06704174896207231, "p90": 0.19211866478328898, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 14.47086411264354, "p90": 31.61392907921375, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0978271686446571, "p90": 0.28246504738708383, "valid_pairs": 562, "valid_duration_s": 18.73333333333333, "coverage": 0.9382303839732887 }, "cmh_m_s": { "median": 0.061186941215898974, "p90": 0.16642789317781484, "valid_pairs": 562, "valid_duration_s": 18.73333333333333, "coverage": 0.9382303839732887 }, "rhcm_continuous_estimate_m_s": { "median": 0.12783666436659413, "p90": 0.34607076058529535, "valid_pairs": 562, "valid_duration_s": 18.73333333333333, "coverage": 0.9382303839732887 }, "rhcm_discrete_m_s": { "median": 0.12759679854321154, "p90": 0.3467585562894852, "valid_pairs": 562, "valid_duration_s": 18.73333333333333, "coverage": 0.9382303839732887 }, "wms": { "median": 0.7871263603307168, "p90": 0.9742295636297575, "valid_pairs": 561, "valid_duration_s": 18.699999999999996, "coverage": 0.9365609348914856 }, "rham_deg_s": { "median": 46.22296377902492, "p90": 119.84025278471351, "valid_pairs": 562, "valid_duration_s": 18.73333333333333, "coverage": 0.9382303839732887 }, "speed_spearman": 0.08603212396068262, "hand_class": "low", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.10026890039256503, "p90": 0.3092975482145302, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549249 }, "cmh_m_s": { "median": 0.07728488391781445, "p90": 0.1763907337033537, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549249 }, "rhcm_continuous_estimate_m_s": { "median": 0.13399150616536948, "p90": 0.30594969961877816, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549249 }, "rhcm_discrete_m_s": { "median": 0.13364342685601113, "p90": 0.3051649857975632, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549249 }, "wms": { "median": 0.7479707299555737, "p90": 0.9591074733119394, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549249 }, "rham_deg_s": { "median": 46.068780640690946, "p90": 138.30832205651487, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549249 }, "speed_spearman": 0.1439682548495349, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "mixed", "relative_both_class": "high", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.19211866478328898, "camera_angular_speed_p90_deg_s": 31.61392907921375, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3092975482145302, "hand_camera_speed_p90_m_s": 0.3467585562894852, "hand_world_valid_fraction": 0.9483333333333334, "hand_camera_valid_fraction": 0.9483333333333334, "cmh_p90_m_s": 0.1763907337033537, "rhcm_p90_m_s": 0.3467585562894852, "rham_p90_deg_s": 138.30832205651487, "wms_p90": 0.9742295636297575, "gt_coverage": 0.9582637729549249, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "0cfbcffc75eec3daf553088eb0599c2d19c1097d19637dea1b7e3c2c63790f5d", "metadata_sha256": "a71313d8398cecccceab9efb642908af17cfd912255a504a68542957340c1bbd", "source_file": "long20_20260918/local/benchmark/previews/6644e508bb0fdf63e6777f28/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.652682298944613e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b7550c690f7f2614eee48f65036d0ebca609ecf3878e51e67a5cf7ce0f65a22a", "mano_sha256": "770d77711006c6cdd1ccac2b62b378de9e9d1ef4adf37fce621fd5e340904ed6", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "6c94bb9c47ec275ec5973f6e", "sample_id": "0be4ef3ada6f35b0d9290ea1", "file": "20_seconds/ARCTIC/0be4ef3ada6f35b0d9290ea1_MANO_GT_20.00s.mp4", "recording_id": "s05/scissors_use_02", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 23/600 output frames; may include genuine out-of-view frames", "id": "6c94bb9c47ec275ec5973f6e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/0be4ef3ada6f35b0d9290ea1_MANO_GT_20.00s.mp4", "bytes": 4628502, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.062296304044039665, "p90": 0.1688813924168105, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 9.771465496383183, "p90": 30.954053069404136, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.05200670187158819, "p90": 0.20769302610705462, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "cmh_m_s": { "median": 0.0625486410699897, "p90": 0.16891830673296196, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "rhcm_continuous_estimate_m_s": { "median": 0.07469178559134462, "p90": 0.22377506505262726, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "rhcm_discrete_m_s": { "median": 0.07487100333863292, "p90": 0.22425922454775474, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "wms": { "median": 0.6348148100906501, "p90": 0.9286106356722434, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "rham_deg_s": { "median": 27.58412529280448, "p90": 113.51226219925442, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "speed_spearman": 0.3963547917665837, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.03841697895974297, "p90": 0.1641446551587969, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "cmh_m_s": { "median": 0.057240931122380205, "p90": 0.14484932111615567, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "rhcm_continuous_estimate_m_s": { "median": 0.06479726598783836, "p90": 0.2018869926588762, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "rhcm_discrete_m_s": { "median": 0.06488497572865148, "p90": 0.20121100936159797, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "wms": { "median": 0.6493711324421931, "p90": 0.912915967367334, "valid_pairs": 508, "valid_duration_s": 16.93333333333333, "coverage": 0.8480801335559265 }, "rham_deg_s": { "median": 20.144249714909577, "p90": 77.71085009711211, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "speed_spearman": 0.39307500177904214, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1688813924168105, "camera_angular_speed_p90_deg_s": 30.954053069404136, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.20769302610705462, "hand_camera_speed_p90_m_s": 0.22425922454775474, "hand_world_valid_fraction": 0.885, "hand_camera_valid_fraction": 0.885, "cmh_p90_m_s": 0.16891830673296196, "rhcm_p90_m_s": 0.22425922454775474, "rham_p90_deg_s": 113.51226219925442, "wms_p90": 0.9286106356722434, "gt_coverage": 0.9048414023372288, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "093175d9b44f6065a47b15af65aa4294dfcf33be5b80a2bcb7846b89a5067302", "metadata_sha256": "b74cb1d9bb476944a9c7c94fe448f9b96c0ef7ee095131cb65a5608a0e9b1f18", "source_file": "long20_20260918/local/benchmark/previews/6c94bb9c47ec275ec5973f6e/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.611102534756203e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "86d66767c8d0490f37ef3366b3583b06f4c7d0491f02c30553aed03a95e39313", "mano_sha256": "84ecdafdc24ed496594876ad10d3e2a8f4a1f0d6d12af9f812618f3f72f16fd0", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "7514cec1e4828da2526ae53c", "sample_id": "eed1c3cd035c4b05c6ebde8c", "file": "20_seconds/ARCTIC/eed1c3cd035c4b05c6ebde8c_MANO_GT_20.00s.mp4", "recording_id": "s05/phone_use_03", "source_first_frame": "226", "source_last_frame": "825", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 21/600 output frames; may include genuine out-of-view frames", "id": "7514cec1e4828da2526ae53c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/eed1c3cd035c4b05c6ebde8c_MANO_GT_20.00s.mp4", "bytes": 5737816, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.04137794416280024, "p90": 0.135857620115109, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 10.584547614717613, "p90": 30.01289097110656, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08414849346238537, "p90": 0.1868451828507854, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "cmh_m_s": { "median": 0.054990440459600574, "p90": 0.12768824832792125, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "rhcm_continuous_estimate_m_s": { "median": 0.0974687144894799, "p90": 0.21580749335260693, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "rhcm_discrete_m_s": { "median": 0.09743258008948935, "p90": 0.2157099305401502, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "wms": { "median": 0.7589161417921001, "p90": 0.9586895449796264, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "rham_deg_s": { "median": 62.35350999783275, "p90": 162.30718154530987, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "speed_spearman": 0.35152782945546596, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.06046687631510234, "p90": 0.12326646605543568, "valid_pairs": 469, "valid_duration_s": 15.633333333333333, "coverage": 0.7829716193656094 }, "cmh_m_s": { "median": 0.05419427717740207, "p90": 0.14002073287693093, "valid_pairs": 469, "valid_duration_s": 15.633333333333333, "coverage": 0.7829716193656094 }, "rhcm_continuous_estimate_m_s": { "median": 0.08696450450894555, "p90": 0.1606579978752765, "valid_pairs": 469, "valid_duration_s": 15.633333333333333, "coverage": 0.7829716193656094 }, "rhcm_discrete_m_s": { "median": 0.08704122228859142, "p90": 0.16019711219602703, "valid_pairs": 469, "valid_duration_s": 15.633333333333333, "coverage": 0.7829716193656094 }, "wms": { "median": 0.7110757837895143, "p90": 0.9134451868790859, "valid_pairs": 468, "valid_duration_s": 15.599999999999998, "coverage": 0.7813021702838063 }, "rham_deg_s": { "median": 61.767192066869136, "p90": 179.99947842151158, "valid_pairs": 469, "valid_duration_s": 15.633333333333333, "coverage": 0.7829716193656094 }, "speed_spearman": 0.23579146586450497, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.135857620115109, "camera_angular_speed_p90_deg_s": 30.01289097110656, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.1868451828507854, "hand_camera_speed_p90_m_s": 0.2157099305401502, "hand_world_valid_fraction": 0.8691666666666666, "hand_camera_valid_fraction": 0.8691666666666666, "cmh_p90_m_s": 0.14002073287693093, "rhcm_p90_m_s": 0.2157099305401502, "rham_p90_deg_s": 179.99947842151158, "wms_p90": 0.9586895449796264, "gt_coverage": 0.9465776293823038, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "b8a55fdca6630c61cffe72ce2085a3fdf5520c5b9a68b47c806a865601ba36b8", "metadata_sha256": "7d51a54f2de1d06bdc6f22fbfaec314ef10018bcb0e048564a811712b3d8c3a1", "source_file": "long20_20260918/local/benchmark/previews/7514cec1e4828da2526ae53c/reference.npz", "window_s": [ 7.5, 27.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.509747244882934e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "069c9dade248c9a162cd197531369aaabe90a340b5c62c880ac7fda747f1a2e2", "mano_sha256": "676e07ad60ec4945756194151ff94b8598b3726cb6446ef11d865c6c0eb22d69", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "7543ae9bc44fe472414e00ab", "sample_id": "d06d04696786692b56f831b8", "file": "20_seconds/ARCTIC/d06d04696786692b56f831b8_MANO_GT_20.00s.mp4", "recording_id": "s05/scissors_use_01_retake", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 26/600 output frames; may include genuine out-of-view frames", "id": "7543ae9bc44fe472414e00ab", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/d06d04696786692b56f831b8_MANO_GT_20.00s.mp4", "bytes": 4853928, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.06313168263819234, "p90": 0.16544867793506157, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 9.642138170494789, "p90": 28.21050381803224, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07479212163651949, "p90": 0.24361689855024207, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "cmh_m_s": { "median": 0.07241064364204726, "p90": 0.18823387354287577, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "rhcm_continuous_estimate_m_s": { "median": 0.09797702735506957, "p90": 0.26389012813293133, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "rhcm_discrete_m_s": { "median": 0.09826261611777494, "p90": 0.2632100309487747, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "wms": { "median": 0.6594590202686066, "p90": 0.8905953724018949, "valid_pairs": 544, "valid_duration_s": 18.13333333333333, "coverage": 0.9081803005008346 }, "rham_deg_s": { "median": 38.24407616717893, "p90": 114.52457276257032, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "speed_spearman": 0.5340609396586422, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.07847515684600562, "p90": 0.2081203046260847, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 0.7312186978297163 }, "cmh_m_s": { "median": 0.08880886794328907, "p90": 0.22176024960945892, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 0.7312186978297163 }, "rhcm_continuous_estimate_m_s": { "median": 0.10929021753795973, "p90": 0.2897344035406809, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 0.7312186978297163 }, "rhcm_discrete_m_s": { "median": 0.1092691999511978, "p90": 0.2897968093496181, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 0.7312186978297163 }, "wms": { "median": 0.6238162637665248, "p90": 0.9379362289728029, "valid_pairs": 432, "valid_duration_s": 14.399999999999997, "coverage": 0.7212020033388981 }, "rham_deg_s": { "median": 27.848153845179528, "p90": 128.54042382522684, "valid_pairs": 438, "valid_duration_s": 14.6, "coverage": 0.7312186978297163 }, "speed_spearman": 0.592012132637854, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.16544867793506157, "camera_angular_speed_p90_deg_s": 28.21050381803224, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.24361689855024207, "hand_camera_speed_p90_m_s": 0.2897968093496181, "hand_world_valid_fraction": 0.8275, "hand_camera_valid_fraction": 0.8275, "cmh_p90_m_s": 0.22176024960945892, "rhcm_p90_m_s": 0.2897968093496181, "rham_p90_deg_s": 128.54042382522684, "wms_p90": 0.9379362289728029, "gt_coverage": 0.9165275459098496, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "6fb7409e5436808d5a3842ef7e1a37a20a89a2daac6ac4e8e15891ca7b8509f3", "metadata_sha256": "5f23af475d597580a0ebacaaba1f5759518750c89c589fcd814ecbc30d37a14f", "source_file": "long20_20260918/local/benchmark/previews/7543ae9bc44fe472414e00ab/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.554367738890935e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6d1602336b7bd6be14759a7d1dcec5a7fa02104b3b392ac468e8d9c8f593d5f2", "mano_sha256": "7257b3d69a8e5b73c32432d8d5363cabae3f961f877d66d231be54a6438dce8b", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "77f1c8c01815b1ad73d2395d", "sample_id": "77d72bd94e4f0a6e0fdd7f4d", "file": "20_seconds/ARCTIC/77d72bd94e4f0a6e0fdd7f4d_MANO_GT_20.00s.mp4", "recording_id": "s05/notebook_grab_01", "source_first_frame": "343", "source_last_frame": "942", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 84/600 output frames; may include genuine out-of-view frames", "id": "77f1c8c01815b1ad73d2395d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/77d72bd94e4f0a6e0fdd7f4d_MANO_GT_20.00s.mp4", "bytes": 5379648, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.07580172222442222, "p90": 0.19340136719677206, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 15.61722167144375, "p90": 40.91142145914961, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.2128723185210011, "p90": 0.48515003017090047, "valid_pairs": 316, "valid_duration_s": 10.53333333333333, "coverage": 0.5275459098497494 }, "cmh_m_s": { "median": 0.08667182662570559, "p90": 0.24180813059045758, "valid_pairs": 316, "valid_duration_s": 10.53333333333333, "coverage": 0.5275459098497494 }, "rhcm_continuous_estimate_m_s": { "median": 0.2240328714278429, "p90": 0.5018855893291828, "valid_pairs": 316, "valid_duration_s": 10.53333333333333, "coverage": 0.5275459098497494 }, "rhcm_discrete_m_s": { "median": 0.22501760851954813, "p90": 0.5024016948096737, "valid_pairs": 316, "valid_duration_s": 10.53333333333333, "coverage": 0.5275459098497494 }, "wms": { "median": 0.7451261132640783, "p90": 0.9360852209772479, "valid_pairs": 316, "valid_duration_s": 10.53333333333333, "coverage": 0.5275459098497494 }, "rham_deg_s": { "median": 81.10047498049477, "p90": 237.55667625529807, "valid_pairs": 316, "valid_duration_s": 10.53333333333333, "coverage": 0.5275459098497494 }, "speed_spearman": 0.3677849337235194, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.2073344440991669, "p90": 0.44521487905889456, "valid_pairs": 476, "valid_duration_s": 15.866666666666664, "coverage": 0.7946577629382302 }, "cmh_m_s": { "median": 0.10185191557510291, "p90": 0.24758040891494265, "valid_pairs": 476, "valid_duration_s": 15.866666666666664, "coverage": 0.7946577629382302 }, "rhcm_continuous_estimate_m_s": { "median": 0.23666188372623104, "p90": 0.49888989952433804, "valid_pairs": 476, "valid_duration_s": 15.866666666666664, "coverage": 0.7946577629382302 }, "rhcm_discrete_m_s": { "median": 0.23677352054325196, "p90": 0.4983439469368824, "valid_pairs": 476, "valid_duration_s": 15.866666666666664, "coverage": 0.7946577629382302 }, "wms": { "median": 0.7615614002311905, "p90": 0.9499034204269146, "valid_pairs": 476, "valid_duration_s": 15.866666666666664, "coverage": 0.7946577629382302 }, "rham_deg_s": { "median": 91.61244419025132, "p90": 237.74500663540476, "valid_pairs": 476, "valid_duration_s": 15.866666666666664, "coverage": 0.7946577629382302 }, "speed_spearman": 0.2355575954314198, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.19340136719677206, "camera_angular_speed_p90_deg_s": 40.91142145914961, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.48515003017090047, "hand_camera_speed_p90_m_s": 0.5024016948096737, "hand_world_valid_fraction": 0.6683333333333333, "hand_camera_valid_fraction": 0.6683333333333333, "cmh_p90_m_s": 0.24758040891494265, "rhcm_p90_m_s": 0.5024016948096737, "rham_p90_deg_s": 237.74500663540476, "wms_p90": 0.9499034204269146, "gt_coverage": 0.7946577629382302, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "13ae69e4976dfc07f08c184fb305261fb5dad547c8c87de7db8d12bc5040eb0c", "metadata_sha256": "85f7f3a0b3769f98bb75148837b45a8de39fadc5e17b5a349f0f78b3c8180383", "source_file": "long20_20260918/local/benchmark/previews/77f1c8c01815b1ad73d2395d/reference.npz", "window_s": [ 11.4, 31.366666666666667 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.671272446551088e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e4a2eba7040700c667a487da4b6b182533f7c09a153404841b5fab50e92801c1", "mano_sha256": "d333ff4c5719f6ecd18a571c44e395f855d26da899143af8b479df94a82728ed", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "78c0bda9a3b240c38ca12e64", "sample_id": "4db15bde95764639bed7395b", "file": "20_seconds/ARCTIC/4db15bde95764639bed7395b_MANO_GT_20.00s.mp4", "recording_id": "s05/scissors_use_01_retake", "source_first_frame": "241", "source_last_frame": "840", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 61/600 output frames; may include genuine out-of-view frames", "id": "78c0bda9a3b240c38ca12e64", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/4db15bde95764639bed7395b_MANO_GT_20.00s.mp4", "bytes": 4655603, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05155757619776716, "p90": 0.14983657789502108, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 7.414407852848904, "p90": 25.321244177319024, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06791679206750117, "p90": 0.21004391780941484, "valid_pairs": 538, "valid_duration_s": 17.933333333333334, "coverage": 0.8981636060100168 }, "cmh_m_s": { "median": 0.07016140288397665, "p90": 0.18344859937415176, "valid_pairs": 538, "valid_duration_s": 17.933333333333334, "coverage": 0.8981636060100168 }, "rhcm_continuous_estimate_m_s": { "median": 0.0920508628570226, "p90": 0.2511507069027705, "valid_pairs": 538, "valid_duration_s": 17.933333333333334, "coverage": 0.8981636060100168 }, "rhcm_discrete_m_s": { "median": 0.09206321649068298, "p90": 0.25080768837000134, "valid_pairs": 538, "valid_duration_s": 17.933333333333334, "coverage": 0.8981636060100168 }, "wms": { "median": 0.6476423711838225, "p90": 0.8901479180924576, "valid_pairs": 534, "valid_duration_s": 17.799999999999997, "coverage": 0.8914858096828046 }, "rham_deg_s": { "median": 36.31202690615964, "p90": 107.394859240936, "valid_pairs": 538, "valid_duration_s": 17.933333333333334, "coverage": 0.8981636060100168 }, "speed_spearman": 0.577996307148943, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06945273291896348, "p90": 0.19466393093657344, "valid_pairs": 498, "valid_duration_s": 16.6, "coverage": 0.8313856427378966 }, "cmh_m_s": { "median": 0.07218062666709135, "p90": 0.20229743739706238, "valid_pairs": 498, "valid_duration_s": 16.6, "coverage": 0.8313856427378966 }, "rhcm_continuous_estimate_m_s": { "median": 0.09672254776787137, "p90": 0.24719126542527967, "valid_pairs": 498, "valid_duration_s": 16.6, "coverage": 0.8313856427378966 }, "rhcm_discrete_m_s": { "median": 0.09682176661480846, "p90": 0.2468793036453822, "valid_pairs": 498, "valid_duration_s": 16.6, "coverage": 0.8313856427378966 }, "wms": { "median": 0.6381435242306681, "p90": 0.9486444140831437, "valid_pairs": 492, "valid_duration_s": 16.4, "coverage": 0.8213689482470785 }, "rham_deg_s": { "median": 29.72057162301155, "p90": 136.79288261888175, "valid_pairs": 498, "valid_duration_s": 16.6, "coverage": 0.8313856427378966 }, "speed_spearman": 0.6188194186729864, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.14983657789502108, "camera_angular_speed_p90_deg_s": 25.321244177319024, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.21004391780941484, "hand_camera_speed_p90_m_s": 0.25080768837000134, "hand_world_valid_fraction": 0.8658333333333333, "hand_camera_valid_fraction": 0.8658333333333333, "cmh_p90_m_s": 0.20229743739706238, "rhcm_p90_m_s": 0.25080768837000134, "rham_p90_deg_s": 136.79288261888175, "wms_p90": 0.9486444140831437, "gt_coverage": 0.8981636060100168, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "5a40e24b34a9f599ed36dc13a3e76aa17e95eef74a82f222f0daffa193a17c9c", "metadata_sha256": "5f23af475d597580a0ebacaaba1f5759518750c89c589fcd814ecbc30d37a14f", "source_file": "long20_20260918/local/benchmark/previews/78c0bda9a3b240c38ca12e64/reference.npz", "window_s": [ 8.0, 27.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.671272446551088e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2a44a280f746706378941ae9bde54d4438ebd6aff6c6ee0866013373edba2d17", "mano_sha256": "57b8d1618042eb1939f92ff5fea2470677181edf94b0c583b01a30883f955563", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "790823e7d658e53c17d2e477", "sample_id": "73a9ea2e400ae1e2e67f7e1d", "file": "20_seconds/ARCTIC/73a9ea2e400ae1e2e67f7e1d_MANO_GT_20.00s.mp4", "recording_id": "s05/laptop_use_02", "source_first_frame": "121", "source_last_frame": "720", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 50/600 output frames; may include genuine out-of-view frames", "id": "790823e7d658e53c17d2e477", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/73a9ea2e400ae1e2e67f7e1d_MANO_GT_20.00s.mp4", "bytes": 5692627, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08742974692280547, "p90": 0.23140182917442798, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 16.65995653404154, "p90": 41.737730480303895, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06810532361326858, "p90": 0.34776103830586863, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "cmh_m_s": { "median": 0.12642927382534846, "p90": 0.27299420003267444, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "rhcm_continuous_estimate_m_s": { "median": 0.14315483116346883, "p90": 0.4501929756600583, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "rhcm_discrete_m_s": { "median": 0.14309983039698942, "p90": 0.4481776971868875, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "wms": { "median": 0.8155199508379476, "p90": 0.967861360095887, "valid_pairs": 541, "valid_duration_s": 18.03333333333333, "coverage": 0.9031719532554257 }, "rham_deg_s": { "median": 41.43671517915278, "p90": 115.0357085294149, "valid_pairs": 542, "valid_duration_s": 18.066666666666666, "coverage": 0.9048414023372288 }, "speed_spearman": 0.4505221865159695, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.21751681188155955, "p90": 0.5837320882606076, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193657 }, "cmh_m_s": { "median": 0.12031599188137661, "p90": 0.2915720742306519, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193657 }, "rhcm_continuous_estimate_m_s": { "median": 0.22240652640551456, "p90": 0.5775794832398986, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193657 }, "rhcm_discrete_m_s": { "median": 0.2227474592760026, "p90": 0.5782789547040799, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193657 }, "wms": { "median": 0.6695359195474629, "p90": 0.9437360238138197, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193657 }, "rham_deg_s": { "median": 64.30874516404468, "p90": 145.53212319789603, "valid_pairs": 521, "valid_duration_s": 17.366666666666667, "coverage": 0.8697829716193657 }, "speed_spearman": 0.2803071186981439, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.23140182917442798, "camera_angular_speed_p90_deg_s": 41.737730480303895, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5837320882606076, "hand_camera_speed_p90_m_s": 0.5782789547040799, "hand_world_valid_fraction": 0.8891666666666667, "hand_camera_valid_fraction": 0.8891666666666667, "cmh_p90_m_s": 0.2915720742306519, "rhcm_p90_m_s": 0.5782789547040799, "rham_p90_deg_s": 145.53212319789603, "wms_p90": 0.967861360095887, "gt_coverage": 0.9048414023372288, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "33dc893e56d6e73663c941de703af441e2b965a0a0619a13607f6b530dff8241", "metadata_sha256": "fbbbe89655db0409b0f53144d61f0acda364365e5b8097dd6fb661143e56c56c", "source_file": "long20_20260918/local/benchmark/previews/790823e7d658e53c17d2e477/reference.npz", "window_s": [ 4.0, 23.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.509747244882934e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2bffd38f7d5223a117697aa2ccfffc96f7d5a24efb4256ac716fcd58e7938eff", "mano_sha256": "3eec45e7bfa1c5e1cf90d90aeb118703c965880dde789571d578511797b05957", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "7b57754397b6b6a3671611c3", "sample_id": "b1c2ca7a4ccf217a285ecf3d", "file": "20_seconds/ARCTIC/b1c2ca7a4ccf217a285ecf3d_MANO_GT_20.00s.mp4", "recording_id": "s08/capsulemachine_use_03", "source_first_frame": "2", "source_last_frame": "601", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 49/600 output frames; may include genuine out-of-view frames", "id": "7b57754397b6b6a3671611c3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/b1c2ca7a4ccf217a285ecf3d_MANO_GT_20.00s.mp4", "bytes": 4638268, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.019456755898854528, "p90": 0.048066802658978905, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 4.436652746675451, "p90": 8.893188388771998, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.09151232356995354, "p90": 0.26699543902224415, "valid_pairs": 505, "valid_duration_s": 16.833333333333336, "coverage": 0.8430717863105177 }, "cmh_m_s": { "median": 0.026533479880368582, "p90": 0.05352693812290498, "valid_pairs": 505, "valid_duration_s": 16.833333333333336, "coverage": 0.8430717863105177 }, "rhcm_continuous_estimate_m_s": { "median": 0.09951396661901375, "p90": 0.26825771293409695, "valid_pairs": 505, "valid_duration_s": 16.833333333333336, "coverage": 0.8430717863105177 }, "rhcm_discrete_m_s": { "median": 0.0995207412720685, "p90": 0.2682660443814368, "valid_pairs": 505, "valid_duration_s": 16.833333333333336, "coverage": 0.8430717863105177 }, "wms": { "median": 0.8519358546918143, "p90": 0.9692190354453872, "valid_pairs": 492, "valid_duration_s": 16.400000000000002, "coverage": 0.8213689482470786 }, "rham_deg_s": { "median": 48.01024419068584, "p90": 144.73146903545793, "valid_pairs": 505, "valid_duration_s": 16.833333333333336, "coverage": 0.8430717863105177 }, "speed_spearman": -0.04157437482878724, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09557523705311934, "p90": 0.22608735398889648, "valid_pairs": 550, "valid_duration_s": 18.333333333333332, "coverage": 0.9181969949916527 }, "cmh_m_s": { "median": 0.027536095640764327, "p90": 0.05583803104686869, "valid_pairs": 550, "valid_duration_s": 18.333333333333332, "coverage": 0.9181969949916527 }, "rhcm_continuous_estimate_m_s": { "median": 0.09563766440091995, "p90": 0.23850865474207997, "valid_pairs": 550, "valid_duration_s": 18.333333333333332, "coverage": 0.9181969949916527 }, "rhcm_discrete_m_s": { "median": 0.09569181416847304, "p90": 0.2385088616845525, "valid_pairs": 550, "valid_duration_s": 18.333333333333332, "coverage": 0.9181969949916527 }, "wms": { "median": 0.8457022114609898, "p90": 0.9616818150225367, "valid_pairs": 547, "valid_duration_s": 18.233333333333327, "coverage": 0.9131886477462435 }, "rham_deg_s": { "median": 48.41224713627898, "p90": 146.65071061293253, "valid_pairs": 550, "valid_duration_s": 18.333333333333332, "coverage": 0.9181969949916527 }, "speed_spearman": 0.20683968001096323, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.048066802658978905, "camera_angular_speed_p90_deg_s": 8.893188388771998, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.26699543902224415, "hand_camera_speed_p90_m_s": 0.2682660443814368, "hand_world_valid_fraction": 0.8816666666666667, "hand_camera_valid_fraction": 0.8816666666666667, "cmh_p90_m_s": 0.05583803104686869, "rhcm_p90_m_s": 0.2682660443814368, "rham_p90_deg_s": 146.65071061293253, "wms_p90": 0.9692190354453872, "gt_coverage": 0.9181969949916527, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "def1e563b56b0d0177ce4c16fb9cd4ff0c7b2f5c89059643fd6db20b2d298635", "metadata_sha256": "2152cc28dd758f0541f574c2bf646e29fc7ef8649d93f4c0394e7d187c3c9ece", "source_file": "long20_20260918/local/train/previews/7b57754397b6b6a3671611c3/reference.npz", "window_s": [ 0.03333333333333333, 20.0 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.585973723508303e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5251699f839c5e2e87f3648e571b0b252b5fefab737566425a11ef3ec4a68180", "mano_sha256": "c1d9cd6811a152a11510f51f466fc84422c7990c22c5b3cb5653725beab0791e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "7b6392a61d6992107cd364df", "sample_id": "b3a715ad735626736bd715cd", "file": "20_seconds/ARCTIC/b3a715ad735626736bd715cd_MANO_GT_20.00s.mp4", "recording_id": "s05/phone_use_04", "source_first_frame": "517", "source_last_frame": "1116", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 63/600 output frames; may include genuine out-of-view frames", "id": "7b6392a61d6992107cd364df", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/b3a715ad735626736bd715cd_MANO_GT_20.00s.mp4", "bytes": 5608492, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.06496590037103211, "p90": 0.12659891815927204, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 15.485103418443707, "p90": 31.269887377224247, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.071098434380862, "p90": 0.19340490957328202, "valid_pairs": 522, "valid_duration_s": 17.400000000000006, "coverage": 0.871452420701169 }, "cmh_m_s": { "median": 0.05654974752593199, "p90": 0.1289150628619898, "valid_pairs": 522, "valid_duration_s": 17.400000000000006, "coverage": 0.871452420701169 }, "rhcm_continuous_estimate_m_s": { "median": 0.0998727631901934, "p90": 0.2513248673581392, "valid_pairs": 522, "valid_duration_s": 17.400000000000006, "coverage": 0.871452420701169 }, "rhcm_discrete_m_s": { "median": 0.09998292938905277, "p90": 0.2511060065570877, "valid_pairs": 522, "valid_duration_s": 17.400000000000006, "coverage": 0.871452420701169 }, "wms": { "median": 0.7772856885549134, "p90": 0.9573675904589931, "valid_pairs": 522, "valid_duration_s": 17.400000000000006, "coverage": 0.871452420701169 }, "rham_deg_s": { "median": 51.81975260907903, "p90": 126.74107426786716, "valid_pairs": 522, "valid_duration_s": 17.400000000000006, "coverage": 0.871452420701169 }, "speed_spearman": 0.1589605864265454, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06715312247085838, "p90": 0.2603372520968059, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113524 }, "cmh_m_s": { "median": 0.06562985323033493, "p90": 0.13809639756820716, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113524 }, "rhcm_continuous_estimate_m_s": { "median": 0.10801863535922845, "p90": 0.24723772310296602, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113524 }, "rhcm_discrete_m_s": { "median": 0.10774539697164709, "p90": 0.24531543634266928, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113524 }, "wms": { "median": 0.752985011725325, "p90": 0.9270883023717755, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113524 }, "rham_deg_s": { "median": 52.88773473054694, "p90": 147.72399232751457, "valid_pairs": 450, "valid_duration_s": 15.0, "coverage": 0.7512520868113524 }, "speed_spearman": 0.08652730795378413, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.12659891815927204, "camera_angular_speed_p90_deg_s": 31.269887377224247, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2603372520968059, "hand_camera_speed_p90_m_s": 0.2511060065570877, "hand_world_valid_fraction": 0.8166666666666667, "hand_camera_valid_fraction": 0.8166666666666667, "cmh_p90_m_s": 0.13809639756820716, "rhcm_p90_m_s": 0.2511060065570877, "rham_p90_deg_s": 147.72399232751457, "wms_p90": 0.9573675904589931, "gt_coverage": 0.871452420701169, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "09cdd8c350eb34f43fc586ba7308eb07b1dd158a72c6925a59cee3ad67bd53c6", "metadata_sha256": "51a55642c92277ea8de693c5e7e48ab093f68d7968c3a1444b9bdbc2a96f9404", "source_file": "long20_20260918/local/benchmark/previews/7b6392a61d6992107cd364df/reference.npz", "window_s": [ 17.2, 37.166666666666664 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.660953740672398e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "fe70d4a90ee9e61993ea47ee22f92f8a06883d14505c1c521de1f8ffbe42f943", "mano_sha256": "823dbe7d4478aaeffd4717c01f4c99984f437d29dd2157f16ed3ef528859d649", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "7bf9e32a6c6565a69b0347e7", "sample_id": "0821cffa3e4eff7b3a3c92f2", "file": "20_seconds/ARCTIC/0821cffa3e4eff7b3a3c92f2_MANO_GT_20.00s.mp4", "recording_id": "s07/scissors_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 21/600 output frames; may include genuine out-of-view frames", "id": "7bf9e32a6c6565a69b0347e7", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/0821cffa3e4eff7b3a3c92f2_MANO_GT_20.00s.mp4", "bytes": 4574547, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.029061481758998142, "p90": 0.09854037906995002, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 5.172319448772077, "p90": 18.189367340303413, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.04596335649870356, "p90": 0.1540738652718785, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "cmh_m_s": { "median": 0.033292924199337486, "p90": 0.07861969020043068, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "rhcm_continuous_estimate_m_s": { "median": 0.04354206212563293, "p90": 0.11265519888793263, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "rhcm_discrete_m_s": { "median": 0.04353907390283622, "p90": 0.11167989903238862, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "wms": { "median": 0.5684678763167381, "p90": 0.8836089404772379, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "rham_deg_s": { "median": 19.526542879936695, "p90": 58.72761928182918, "valid_pairs": 539, "valid_duration_s": 17.966666666666665, "coverage": 0.8998330550918197 }, "speed_spearman": 0.48702724077126625, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.04798342634428575, "p90": 0.16860036453737548, "valid_pairs": 578, "valid_duration_s": 19.266666666666666, "coverage": 0.9649415692821369 }, "cmh_m_s": { "median": 0.030679519303507985, "p90": 0.07418452870195802, "valid_pairs": 578, "valid_duration_s": 19.266666666666666, "coverage": 0.9649415692821369 }, "rhcm_continuous_estimate_m_s": { "median": 0.057751672438371315, "p90": 0.15224014426240268, "valid_pairs": 578, "valid_duration_s": 19.266666666666666, "coverage": 0.9649415692821369 }, "rhcm_discrete_m_s": { "median": 0.057761367127814794, "p90": 0.15223739253936297, "valid_pairs": 578, "valid_duration_s": 19.266666666666666, "coverage": 0.9649415692821369 }, "wms": { "median": 0.6499095585319219, "p90": 0.9399010345300729, "valid_pairs": 566, "valid_duration_s": 18.866666666666667, "coverage": 0.9449081803005009 }, "rham_deg_s": { "median": 22.491839311260964, "p90": 103.48818236390437, "valid_pairs": 578, "valid_duration_s": 19.266666666666666, "coverage": 0.9649415692821369 }, "speed_spearman": 0.34557236139244635, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.09854037906995002, "camera_angular_speed_p90_deg_s": 18.189367340303413, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.16860036453737548, "hand_camera_speed_p90_m_s": 0.15223739253936297, "hand_world_valid_fraction": 0.9325, "hand_camera_valid_fraction": 0.9325, "cmh_p90_m_s": 0.07861969020043068, "rhcm_p90_m_s": 0.15223739253936297, "rham_p90_deg_s": 103.48818236390437, "wms_p90": 0.9399010345300729, "gt_coverage": 0.9649415692821369, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "86724584301f00e42758b56637704714fbd2252d1793c31f273b9ad17360909e", "metadata_sha256": "7a046a75078f9b78405f39e445d04cce1f9b81884e64cc037e901d80ca54cf20", "source_file": "long20_20260918/local/train/previews/7bf9e32a6c6565a69b0347e7/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.652682298944613e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "45c402db3fdc3544d748da1716326904b74f5bf01d0017b5c0753886174738cb", "mano_sha256": "3237d1cc5599781601f7ed198f997ae776425aca97cc9df823523e0fc8aa0ed4", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "82ecfcb976f39ae811c3076e", "sample_id": "443cb12650cc023916b8830e", "file": "20_seconds/ARCTIC/443cb12650cc023916b8830e_MANO_GT_20.00s.mp4", "recording_id": "s05/scissors_use_02", "source_first_frame": "88", "source_last_frame": "687", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 58/600 output frames; may include genuine out-of-view frames", "id": "82ecfcb976f39ae811c3076e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/443cb12650cc023916b8830e_MANO_GT_20.00s.mp4", "bytes": 4427268, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05517458778569243, "p90": 0.18263569998069373, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 8.338811771862437, "p90": 28.562594375568235, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.04884654197456579, "p90": 0.1859706372457037, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "cmh_m_s": { "median": 0.06065855645379094, "p90": 0.1721902729375005, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "rhcm_continuous_estimate_m_s": { "median": 0.07076752527589697, "p90": 0.21428900307217516, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "rhcm_discrete_m_s": { "median": 0.07063865383561775, "p90": 0.21590882731455233, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "wms": { "median": 0.6412254491589215, "p90": 0.9293536794200907, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.8881469115191987 }, "rham_deg_s": { "median": 26.632937316403712, "p90": 106.23779154089628, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "speed_spearman": 0.4123019418733928, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.03760146344842479, "p90": 0.1455346467383172, "valid_pairs": 497, "valid_duration_s": 16.566666666666666, "coverage": 0.8297161936560934 }, "cmh_m_s": { "median": 0.053344514384474434, "p90": 0.14127978910861014, "valid_pairs": 497, "valid_duration_s": 16.566666666666666, "coverage": 0.8297161936560934 }, "rhcm_continuous_estimate_m_s": { "median": 0.06234181881887759, "p90": 0.18721835102891007, "valid_pairs": 497, "valid_duration_s": 16.566666666666666, "coverage": 0.8297161936560934 }, "rhcm_discrete_m_s": { "median": 0.06231721960239242, "p90": 0.18668650287430666, "valid_pairs": 497, "valid_duration_s": 16.566666666666666, "coverage": 0.8297161936560934 }, "wms": { "median": 0.6533208041730172, "p90": 0.9039530639696184, "valid_pairs": 490, "valid_duration_s": 16.333333333333336, "coverage": 0.8180300500834725 }, "rham_deg_s": { "median": 19.118617028501973, "p90": 70.72887399162921, "valid_pairs": 497, "valid_duration_s": 16.566666666666666, "coverage": 0.8297161936560934 }, "speed_spearman": 0.3879412307502223, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.18263569998069373, "camera_angular_speed_p90_deg_s": 28.562594375568235, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.1859706372457037, "hand_camera_speed_p90_m_s": 0.21590882731455233, "hand_world_valid_fraction": 0.8683333333333333, "hand_camera_valid_fraction": 0.8683333333333333, "cmh_p90_m_s": 0.1721902729375005, "rhcm_p90_m_s": 0.21590882731455233, "rham_p90_deg_s": 106.23779154089628, "wms_p90": 0.9293536794200907, "gt_coverage": 0.9031719532554257, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "03ae3a5259505b31d580f8c40329f701bdc3dcb52353436e4e5b8617b69923a8", "metadata_sha256": "b74cb1d9bb476944a9c7c94fe448f9b96c0ef7ee095131cb65a5608a0e9b1f18", "source_file": "long20_20260918/local/benchmark/previews/82ecfcb976f39ae811c3076e/reference.npz", "window_s": [ 2.9, 22.866666666666667 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.47545209131181e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "52ff12ac4efa78c34d5922af35786d55d96aab43de279c0e786c79e3aab20317", "mano_sha256": "92febac0704a2e0ee114bc3068ffd19a7b5356f75992a2839413abebe13993e3", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "878590cddc32d6040dbbdb3b", "sample_id": "1f1afc52cae55065ed6e5b57", "file": "20_seconds/ARCTIC/1f1afc52cae55065ed6e5b57_MANO_GT_20.00s.mp4", "recording_id": "s09/ketchup_use_01", "source_first_frame": "2", "source_last_frame": "601", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 67/600 output frames; may include genuine out-of-view frames", "id": "878590cddc32d6040dbbdb3b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/1f1afc52cae55065ed6e5b57_MANO_GT_20.00s.mp4", "bytes": 4498674, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.021280347868788928, "p90": 0.05593052746859625, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 4.166867956802197, "p90": 7.513936248126725, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.10590925279333625, "p90": 0.31411579301350673, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.8881469115191988 }, "cmh_m_s": { "median": 0.03170474273734352, "p90": 0.06738180290001508, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.8881469115191988 }, "rhcm_continuous_estimate_m_s": { "median": 0.11037866794828494, "p90": 0.31502226280096784, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.8881469115191988 }, "rhcm_discrete_m_s": { "median": 0.11043830509666908, "p90": 0.3150266493551594, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.8881469115191988 }, "wms": { "median": 0.8277702632970869, "p90": 0.9686412775318972, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.8881469115191988 }, "rham_deg_s": { "median": 67.90675111245926, "p90": 187.8234639415234, "valid_pairs": 532, "valid_duration_s": 17.733333333333334, "coverage": 0.8881469115191988 }, "speed_spearman": 0.09781648766875374, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.12631057829695572, "p90": 0.38248408379484516, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "cmh_m_s": { "median": 0.032880091629066986, "p90": 0.07043735318528606, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "rhcm_continuous_estimate_m_s": { "median": 0.1322265521559411, "p90": 0.3932979677132393, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "rhcm_discrete_m_s": { "median": 0.13229981811937813, "p90": 0.3933060374768175, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "wms": { "median": 0.8475075948671147, "p90": 0.9738022048971825, "valid_pairs": 527, "valid_duration_s": 17.56666666666667, "coverage": 0.8797996661101839 }, "rham_deg_s": { "median": 56.75608642490183, "p90": 156.85421676677998, "valid_pairs": 528, "valid_duration_s": 17.6, "coverage": 0.8814691151919868 }, "speed_spearman": 0.13645526840197966, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05593052746859625, "camera_angular_speed_p90_deg_s": 7.513936248126725, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.38248408379484516, "hand_camera_speed_p90_m_s": 0.3933060374768175, "hand_world_valid_fraction": 0.885, "hand_camera_valid_fraction": 0.885, "cmh_p90_m_s": 0.07043735318528606, "rhcm_p90_m_s": 0.3933060374768175, "rham_p90_deg_s": 187.8234639415234, "wms_p90": 0.9738022048971825, "gt_coverage": 0.8881469115191988, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "550459fd10c24ba243ca8ded97ca176374a264a3c2d675e55ad56e6642fde0ed", "metadata_sha256": "c2824bad73f5dfabd69cb918514ff34a7de9cdde8092c824950d130d09bca4a3", "source_file": "long20_20260918/local/train/previews/878590cddc32d6040dbbdb3b/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.577566798522237e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2b97a5738ac0c03a47d02ff0a1593ce0416af7e9e9fe8a19dfc268840320dfcb", "mano_sha256": "133198e47bafa1043a74b5abeb6a774aa8c012302ed7d479883501381ba39abe", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "89c35388c235b52bc9ec9eee", "sample_id": "8b7af9310579c88e0f645e51", "file": "20_seconds/ARCTIC/8b7af9310579c88e0f645e51_MANO_GT_20.00s.mp4", "recording_id": "s06/capsulemachine_use_02", "source_first_frame": "28", "source_last_frame": "627", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "", "id": "89c35388c235b52bc9ec9eee", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/8b7af9310579c88e0f645e51_MANO_GT_20.00s.mp4", "bytes": 5304791, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03196292455833339, "p90": 0.09553521578391934, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 5.512440204399842, "p90": 15.943611672711734, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.10256065538494777, "p90": 0.32629960740889774, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0460680986297404, "p90": 0.10934809164385186, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.10928067220337462, "p90": 0.3422198406066167, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.10916775558875837, "p90": 0.34199128657136374, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.7894765679070788, "p90": 0.965528387004388, "valid_pairs": 588, "valid_duration_s": 19.6, "coverage": 0.9816360601001669 }, "rham_deg_s": { "median": 47.90075583583557, "p90": 166.95718631074584, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "speed_spearman": 0.22442778097274718, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.08557842176991541, "p90": 0.24199471972551212, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.044886879931441276, "p90": 0.0997501914715845, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.09933556793130754, "p90": 0.26388680134287623, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.09922462995846647, "p90": 0.2634717732537102, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "wms": { "median": 0.780228662595231, "p90": 0.9617644325463721, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "rham_deg_s": { "median": 47.74046478523731, "p90": 126.54066381796557, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "speed_spearman": -0.0006175844914322085, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.09553521578391934, "camera_angular_speed_p90_deg_s": 15.943611672711734, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.32629960740889774, "hand_camera_speed_p90_m_s": 0.34199128657136374, "hand_world_valid_fraction": 1.0, "hand_camera_valid_fraction": 1.0, "cmh_p90_m_s": 0.10934809164385186, "rhcm_p90_m_s": 0.34199128657136374, "rham_p90_deg_s": 166.95718631074584, "wms_p90": 0.965528387004388, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "fede23c892e51ae742a7a221c30660ec34dc79e9d7cc698b07ac6eae46f2ff8b", "metadata_sha256": "264742cfa38314f78ce4c1d676c10b4703149d37613a21474d793d988bcd3aa0", "source_file": "long20_20260918/local/train/previews/89c35388c235b52bc9ec9eee/reference.npz", "window_s": [ 0.9, 20.866666666666667 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.545902364027965e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5dcad1a917358d35afaa84223aba4df5336414304eefd9943a321cee41578b54", "mano_sha256": "8bc7649f112a2dfd8a6c81661da9a8ed9c7a80aed79bdb2a3e5ae17f075302a2", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "False", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "8a39a95c8e5e731fc6dd69b5", "sample_id": "51b1e3ae9dde18fd293dc238", "file": "20_seconds/ARCTIC/51b1e3ae9dde18fd293dc238_MANO_GT_20.00s.mp4", "recording_id": "s07/notebook_use_02", "source_first_frame": "124", "source_last_frame": "723", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 29/600 output frames; may include genuine out-of-view frames", "id": "8a39a95c8e5e731fc6dd69b5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/51b1e3ae9dde18fd293dc238_MANO_GT_20.00s.mp4", "bytes": 5321840, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05318371841827397, "p90": 0.22521645276902838, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 10.989896203907234, "p90": 45.12151834692623, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.10560677748239361, "p90": 0.33200646233631226, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "cmh_m_s": { "median": 0.06837750456709665, "p90": 0.2304654872294178, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "rhcm_continuous_estimate_m_s": { "median": 0.09214581363533608, "p90": 0.21109266952544573, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "rhcm_discrete_m_s": { "median": 0.09234086769892758, "p90": 0.21103253751403053, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "wms": { "median": 0.5941480800831951, "p90": 0.9218283534205451, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "rham_deg_s": { "median": 39.404636878525025, "p90": 113.75941234090325, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "speed_spearman": 0.49589503199455837, "hand_class": "high", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.06518744119296878, "p90": 0.21748403496417906, "valid_pairs": 143, "valid_duration_s": 4.766666666666668, "coverage": 0.2387312186978298 }, "cmh_m_s": { "median": 0.03744623508062246, "p90": 0.0656118546282582, "valid_pairs": 143, "valid_duration_s": 4.766666666666668, "coverage": 0.2387312186978298 }, "rhcm_continuous_estimate_m_s": { "median": 0.07077317102128838, "p90": 0.2190499918920298, "valid_pairs": 143, "valid_duration_s": 4.766666666666668, "coverage": 0.2387312186978298 }, "rhcm_discrete_m_s": { "median": 0.07079398176446144, "p90": 0.21900248943605286, "valid_pairs": 143, "valid_duration_s": 4.766666666666668, "coverage": 0.2387312186978298 }, "wms": { "median": 0.6992596768848621, "p90": 0.8873014993741886, "valid_pairs": 141, "valid_duration_s": 4.700000000000002, "coverage": 0.2353923205342238 }, "rham_deg_s": { "median": 28.470407847844612, "p90": 88.44355865005961, "valid_pairs": 143, "valid_duration_s": 4.766666666666668, "coverage": 0.2387312186978298 }, "speed_spearman": 0.19629666108539348, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.22521645276902838, "camera_angular_speed_p90_deg_s": 45.12151834692623, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.33200646233631226, "hand_camera_speed_p90_m_s": 0.21900248943605286, "hand_world_valid_fraction": 0.5991666666666666, "hand_camera_valid_fraction": 0.5991666666666666, "cmh_p90_m_s": 0.2304654872294178, "rhcm_p90_m_s": 0.21900248943605286, "rham_p90_deg_s": 113.75941234090325, "wms_p90": 0.9218283534205451, "gt_coverage": 0.9515859766277127, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "3fec259786087260cc3f00fa7a738cf9cc997d1666cee94eb0d87c8e26c74d4c", "metadata_sha256": "45033aaacfc7209c414f50b503636b8d27cf673fbdc63faafdd629a60918cc56", "source_file": "long20_20260918/local/train/previews/8a39a95c8e5e731fc6dd69b5/reference.npz", "window_s": [ 4.1, 24.066666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.849289084079041e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "cbc6fa74e31c5c868832b79fc48ec0cd24d70628a81bdfbce9d4a57125f7cbd5", "mano_sha256": "7e7bb08dab3d53a8f1ba3826450418aa3c2c3dfe5100a4c75960125a64766caf", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "8bf2cc9592bc8a6647b5025c", "sample_id": "321ac2a2363598d05114bcc1", "file": "20_seconds/ARCTIC/321ac2a2363598d05114bcc1_MANO_GT_20.00s.mp4", "recording_id": "s02/ketchup_grab_01", "source_first_frame": "142", "source_last_frame": "741", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 118/600 output frames; may include genuine out-of-view frames", "id": "8bf2cc9592bc8a6647b5025c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/321ac2a2363598d05114bcc1_MANO_GT_20.00s.mp4", "bytes": 3950725, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03128202293407333, "p90": 0.05665010428807555, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 6.851531093546452, "p90": 12.641361389367349, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.2650907347951491, "p90": 0.49910151787009815, "valid_pairs": 478, "valid_duration_s": 15.933333333333337, "coverage": 0.7979966611018366 }, "cmh_m_s": { "median": 0.04425933089325843, "p90": 0.08079046648323707, "valid_pairs": 478, "valid_duration_s": 15.933333333333337, "coverage": 0.7979966611018366 }, "rhcm_continuous_estimate_m_s": { "median": 0.2447666036300442, "p90": 0.47947533628296896, "valid_pairs": 478, "valid_duration_s": 15.933333333333337, "coverage": 0.7979966611018366 }, "rhcm_discrete_m_s": { "median": 0.2448020488092927, "p90": 0.4794459473630764, "valid_pairs": 478, "valid_duration_s": 15.933333333333337, "coverage": 0.7979966611018366 }, "wms": { "median": 0.8050232222710978, "p90": 0.9227373364544544, "valid_pairs": 476, "valid_duration_s": 15.86666666666667, "coverage": 0.7946577629382306 }, "rham_deg_s": { "median": 35.55068399144515, "p90": 74.98749247150928, "valid_pairs": 478, "valid_duration_s": 15.933333333333337, "coverage": 0.7979966611018366 }, "speed_spearman": 0.3449275782710695, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05665010428807555, "camera_angular_speed_p90_deg_s": 12.641361389367349, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.49910151787009815, "hand_camera_speed_p90_m_s": 0.4794459473630764, "hand_world_valid_fraction": 0.40166666666666667, "hand_camera_valid_fraction": 0.40166666666666667, "cmh_p90_m_s": 0.08079046648323707, "rhcm_p90_m_s": 0.4794459473630764, "rham_p90_deg_s": 74.98749247150928, "wms_p90": 0.9227373364544544, "gt_coverage": 0.7979966611018366, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "243045f969b3063cbad1092f5b13f8d939d99419516f7bf2db72c62ac107767e", "metadata_sha256": "14f7e07e6c3903180e22097afd5c0ec32aab1a93ac71246bf112e5c45c693b53", "source_file": "long20_20260918/local/train/previews/8bf2cc9592bc8a6647b5025c/reference.npz", "window_s": [ 4.666666666666667, 24.633333333333333 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.526797389797801e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "fa6dc3229a1baecc1e840b367416d4c9f1e5068afacd0cc2db1282746010104d", "mano_sha256": "236fb6285bc49f41f2ad804186c20a55cb2ce8453c63cae4dd2ae0e2dc08199a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "8d56f7df6aab405c25684e57", "sample_id": "77a8890a35ba7e17447b305c", "file": "20_seconds/ARCTIC/77a8890a35ba7e17447b305c_MANO_GT_20.00s.mp4", "recording_id": "s05/notebook_use_02", "source_first_frame": "76", "source_last_frame": "675", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 15/600 output frames; may include genuine out-of-view frames", "id": "8d56f7df6aab405c25684e57", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/77a8890a35ba7e17447b305c_MANO_GT_20.00s.mp4", "bytes": 6576530, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.1074492561353468, "p90": 0.2096076791223976, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 24.14823191165995, "p90": 47.63758656437627, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.11704062165369805, "p90": 0.33179664093070266, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "cmh_m_s": { "median": 0.11986300801680148, "p90": 0.2793887662411377, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "rhcm_continuous_estimate_m_s": { "median": 0.19688967814113312, "p90": 0.3658147717637842, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "rhcm_discrete_m_s": { "median": 0.19680585781490414, "p90": 0.36589324579945315, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "wms": { "median": 0.7198257043612708, "p90": 0.9082821520163499, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "rham_deg_s": { "median": 68.08092587726283, "p90": 208.7146338492843, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010017 }, "speed_spearman": 0.16428691096588574, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.09988688522998349, "p90": 0.2196970405719822, "valid_pairs": 516, "valid_duration_s": 17.200000000000003, "coverage": 0.8614357262103508 }, "cmh_m_s": { "median": 0.13329650793720232, "p90": 0.27318854457875347, "valid_pairs": 516, "valid_duration_s": 17.200000000000003, "coverage": 0.8614357262103508 }, "rhcm_continuous_estimate_m_s": { "median": 0.1797060952773408, "p90": 0.33140888177543215, "valid_pairs": 516, "valid_duration_s": 17.200000000000003, "coverage": 0.8614357262103508 }, "rhcm_discrete_m_s": { "median": 0.17968288686649075, "p90": 0.3316277591196418, "valid_pairs": 516, "valid_duration_s": 17.200000000000003, "coverage": 0.8614357262103508 }, "wms": { "median": 0.7276409105051458, "p90": 0.9279088496206922, "valid_pairs": 516, "valid_duration_s": 17.200000000000003, "coverage": 0.8614357262103508 }, "rham_deg_s": { "median": 63.11728428460264, "p90": 176.4749096527498, "valid_pairs": 516, "valid_duration_s": 17.200000000000003, "coverage": 0.8614357262103508 }, "speed_spearman": 0.24597542223248997, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.2096076791223976, "camera_angular_speed_p90_deg_s": 47.63758656437627, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.33179664093070266, "hand_camera_speed_p90_m_s": 0.36589324579945315, "hand_world_valid_fraction": 0.8825, "hand_camera_valid_fraction": 0.8825, "cmh_p90_m_s": 0.2793887662411377, "rhcm_p90_m_s": 0.36589324579945315, "rham_p90_deg_s": 208.7146338492843, "wms_p90": 0.9279088496206922, "gt_coverage": 0.8898163606010017, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "e169b5e897a38694ebd525d26d9ab0b7e0b29ffcedbba6f13b3fd3de3c347ba8", "metadata_sha256": "102e5d452d272dd47ef89068336642c7ea536cf15091408ab5cfe720f2d46ec9", "source_file": "long20_20260918/local/benchmark/previews/8d56f7df6aab405c25684e57/reference.npz", "window_s": [ 2.5, 22.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.484050470211362e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7be5c7a61a8000c3834dd1d2d316be22c6e80c2328f64bb639b5d56cacdd177b", "mano_sha256": "de2954991f44595472dd47c707fc492735754d614125a3b941ed170a4e105a81", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "8e406b2ec1eaf9570f79d33b", "sample_id": "b333e0ff852dc258c26532e5", "file": "20_seconds/ARCTIC/b333e0ff852dc258c26532e5_MANO_GT_20.00s.mp4", "recording_id": "s09/scissors_use_01", "source_first_frame": "4", "source_last_frame": "603", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 23/600 output frames; may include genuine out-of-view frames", "id": "8e406b2ec1eaf9570f79d33b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/b333e0ff852dc258c26532e5_MANO_GT_20.00s.mp4", "bytes": 3953224, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.015010503879692503, "p90": 0.04454765478787559, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.686994220812882, "p90": 7.321544284416533, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.0332435793027873, "p90": 0.13839656448970572, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.9565943238731218 }, "cmh_m_s": { "median": 0.022920609399617695, "p90": 0.06225220488605579, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.9565943238731218 }, "rhcm_continuous_estimate_m_s": { "median": 0.03804860711470006, "p90": 0.14157750099640423, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.9565943238731218 }, "rhcm_discrete_m_s": { "median": 0.038049632818116935, "p90": 0.14157553235699857, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.9565943238731218 }, "wms": { "median": 0.6778457840645109, "p90": 0.9198570439651393, "valid_pairs": 561, "valid_duration_s": 18.700000000000003, "coverage": 0.9365609348914858 }, "rham_deg_s": { "median": 20.72160856331453, "p90": 104.7819840629483, "valid_pairs": 573, "valid_duration_s": 19.1, "coverage": 0.9565943238731218 }, "speed_spearman": 0.3287678354081113, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.039158950658533344, "p90": 0.17253225609048484, "valid_pairs": 576, "valid_duration_s": 19.200000000000003, "coverage": 0.9616026711185309 }, "cmh_m_s": { "median": 0.023477172021090578, "p90": 0.061009350745583724, "valid_pairs": 576, "valid_duration_s": 19.200000000000003, "coverage": 0.9616026711185309 }, "rhcm_continuous_estimate_m_s": { "median": 0.04213862466754348, "p90": 0.17425995252053303, "valid_pairs": 576, "valid_duration_s": 19.200000000000003, "coverage": 0.9616026711185309 }, "rhcm_discrete_m_s": { "median": 0.04214351130624381, "p90": 0.17426047531670513, "valid_pairs": 576, "valid_duration_s": 19.200000000000003, "coverage": 0.9616026711185309 }, "wms": { "median": 0.6930886695691667, "p90": 0.954663267515965, "valid_pairs": 572, "valid_duration_s": 19.06666666666667, "coverage": 0.9549248747913189 }, "rham_deg_s": { "median": 23.98078198626489, "p90": 101.71588495993618, "valid_pairs": 576, "valid_duration_s": 19.200000000000003, "coverage": 0.9616026711185309 }, "speed_spearman": 0.19118497726873132, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04454765478787559, "camera_angular_speed_p90_deg_s": 7.321544284416533, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.17253225609048484, "hand_camera_speed_p90_m_s": 0.17426047531670513, "hand_world_valid_fraction": 0.9591666666666666, "hand_camera_valid_fraction": 0.9591666666666666, "cmh_p90_m_s": 0.06225220488605579, "rhcm_p90_m_s": 0.17426047531670513, "rham_p90_deg_s": 104.7819840629483, "wms_p90": 0.954663267515965, "gt_coverage": 0.9616026711185309, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "e23513fe6ae0901e2a32c9536498538df62db9a911d4f277dfabc91b50fb2d8e", "metadata_sha256": "9b66ad1681e91930ba9436a5929f16a8abf0c4a6dc806fe27f9ec8b65e47c49b", "source_file": "long20_20260918/local/train/previews/8e406b2ec1eaf9570f79d33b/reference.npz", "window_s": [ 0.06666666666666667, 20.033333333333335 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.75098925995612e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "37aea27baed09c62982173fe02acae9266a6ad1d0ee6ded74753890f9169bd04", "mano_sha256": "03b653dce35120269e2805f3b9411fc413181cf3aa1596129fefac20fc5108f4", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "8ec5a138d740c91273104cb6", "sample_id": "6454f241cfac74d95b9b07cc", "file": "20_seconds/ARCTIC/6454f241cfac74d95b9b07cc_MANO_GT_20.00s.mp4", "recording_id": "s08/phone_use_03", "source_first_frame": "52", "source_last_frame": "651", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 39/600 output frames; may include genuine out-of-view frames", "id": "8ec5a138d740c91273104cb6", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/6454f241cfac74d95b9b07cc_MANO_GT_20.00s.mp4", "bytes": 4897031, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.024542406401452752, "p90": 0.051324713023139776, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 5.676352977408806, "p90": 14.91084665893399, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.10293743571747964, "p90": 0.2886351944027623, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "cmh_m_s": { "median": 0.03344815161053333, "p90": 0.08773863814720549, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "rhcm_continuous_estimate_m_s": { "median": 0.10159454471747555, "p90": 0.29031317217005137, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "rhcm_discrete_m_s": { "median": 0.10147443322127499, "p90": 0.2903203704372173, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "wms": { "median": 0.7519613835236045, "p90": 0.9620606778477738, "valid_pairs": 510, "valid_duration_s": 17.0, "coverage": 0.8514190317195325 }, "rham_deg_s": { "median": 68.58917543923911, "p90": 178.27229176535593, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "speed_spearman": 0.261503359905606, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08800789389298302, "p90": 0.1908340502475374, "valid_pairs": 560, "valid_duration_s": 18.666666666666664, "coverage": 0.9348914858096826 }, "cmh_m_s": { "median": 0.03396944190745552, "p90": 0.09300901892032461, "valid_pairs": 560, "valid_duration_s": 18.666666666666664, "coverage": 0.9348914858096826 }, "rhcm_continuous_estimate_m_s": { "median": 0.07503608557514968, "p90": 0.17275480437460536, "valid_pairs": 560, "valid_duration_s": 18.666666666666664, "coverage": 0.9348914858096826 }, "rhcm_discrete_m_s": { "median": 0.07504401770781499, "p90": 0.17280680536022192, "valid_pairs": 560, "valid_duration_s": 18.666666666666664, "coverage": 0.9348914858096826 }, "wms": { "median": 0.6531623658416632, "p90": 0.9333448064644141, "valid_pairs": 560, "valid_duration_s": 18.666666666666664, "coverage": 0.9348914858096826 }, "rham_deg_s": { "median": 54.78851781967631, "p90": 138.841505693582, "valid_pairs": 560, "valid_duration_s": 18.666666666666664, "coverage": 0.9348914858096826 }, "speed_spearman": 0.3747120184876684, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.051324713023139776, "camera_angular_speed_p90_deg_s": 14.91084665893399, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2886351944027623, "hand_camera_speed_p90_m_s": 0.2903203704372173, "hand_world_valid_fraction": 0.8966666666666666, "hand_camera_valid_fraction": 0.8966666666666666, "cmh_p90_m_s": 0.09300901892032461, "rhcm_p90_m_s": 0.2903203704372173, "rham_p90_deg_s": 178.27229176535593, "wms_p90": 0.9620606778477738, "gt_coverage": 0.9348914858096826, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "e8e8bd5719925bf2605f700212371ae9f9c1f8ffad02ccb9351ea393135c3763", "metadata_sha256": "ae4a3391a89ee7e22446eb048d2bd66c71fe354e8de82c446b9e43ba72582261", "source_file": "long20_20260918/local/train/previews/8ec5a138d740c91273104cb6/reference.npz", "window_s": [ 1.7, 21.666666666666668 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.552252871008668e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b6a9dcc1eec8de093e207aaebbdb93c1ba450ab9403140771e34daf40b39c77c", "mano_sha256": "84b7a089213fc2a897ccc1c873eb4f83a8cb30a044934502a347eec7aa266873", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "9121d7f9c495a0f0b1367ec5", "sample_id": "ead7ad33ec3dc13a5641b2b1", "file": "20_seconds/ARCTIC/ead7ad33ec3dc13a5641b2b1_MANO_GT_20.00s.mp4", "recording_id": "s10/ketchup_grab_01", "source_first_frame": "36", "source_last_frame": "635", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "", "id": "9121d7f9c495a0f0b1367ec5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/ead7ad33ec3dc13a5641b2b1_MANO_GT_20.00s.mp4", "bytes": 4168591, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.05142988663383648, "p90": 0.11014022123077057, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 8.533400588162687, "p90": 18.283674076272195, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.22800783469034314, "p90": 0.6235627011209252, "valid_pairs": 382, "valid_duration_s": 12.73333333333333, "coverage": 0.6377295492487478 }, "cmh_m_s": { "median": 0.06501751511302095, "p90": 0.15381014941136498, "valid_pairs": 382, "valid_duration_s": 12.73333333333333, "coverage": 0.6377295492487478 }, "rhcm_continuous_estimate_m_s": { "median": 0.2178594848134917, "p90": 0.5928673120496079, "valid_pairs": 382, "valid_duration_s": 12.73333333333333, "coverage": 0.6377295492487478 }, "rhcm_discrete_m_s": { "median": 0.21786577403655427, "p90": 0.5928551053923384, "valid_pairs": 382, "valid_duration_s": 12.73333333333333, "coverage": 0.6377295492487478 }, "wms": { "median": 0.765881970374783, "p90": 0.9556927071832084, "valid_pairs": 381, "valid_duration_s": 12.7, "coverage": 0.6360601001669449 }, "rham_deg_s": { "median": 39.80501871855223, "p90": 85.94449820621863, "valid_pairs": 382, "valid_duration_s": 12.73333333333333, "coverage": 0.6377295492487478 }, "speed_spearman": 0.2731951833020449, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.2141898588292428, "p90": 0.6256312675333657, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 0.6861435726210351 }, "cmh_m_s": { "median": 0.07034542939405444, "p90": 0.16373686213054195, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 0.6861435726210351 }, "rhcm_continuous_estimate_m_s": { "median": 0.19072475086377044, "p90": 0.6629818130820765, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 0.6861435726210351 }, "rhcm_discrete_m_s": { "median": 0.19090431837482544, "p90": 0.6628322664985971, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 0.6861435726210351 }, "wms": { "median": 0.7326793340954952, "p90": 0.9495724868949929, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 0.6861435726210351 }, "rham_deg_s": { "median": 43.03146122402121, "p90": 139.40067975149287, "valid_pairs": 411, "valid_duration_s": 13.7, "coverage": 0.6861435726210351 }, "speed_spearman": 0.2530675774477963, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.11014022123077057, "camera_angular_speed_p90_deg_s": 18.283674076272195, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6256312675333657, "hand_camera_speed_p90_m_s": 0.6628322664985971, "hand_world_valid_fraction": 0.6675, "hand_camera_valid_fraction": 0.6675, "cmh_p90_m_s": 0.16373686213054195, "rhcm_p90_m_s": 0.6628322664985971, "rham_p90_deg_s": 139.40067975149287, "wms_p90": 0.9556927071832084, "gt_coverage": 0.6861435726210351, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "a5d69854415c60b9c39f1a3bfff1063e483ce023a3f07c61e3cf7e6571f61a6c", "metadata_sha256": "b2fd690c9795e956359462709f766ef1687e27e3179f1d60019c9aa6c49017df", "source_file": "long20_20260918/local/train/previews/9121d7f9c495a0f0b1367ec5/reference.npz", "window_s": [ 1.1666666666666667, 21.133333333333333 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 6.106226635438361e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "512081166c6869b2b0150ec86430795d441279908ff56d6f41ad35aa1dd1aa16", "mano_sha256": "e7af8dfb6424cbe4b99103f8b610cdd5444119fc16a0e4edfdf2cac930b50e83", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "92906ad888eef0bf14cecb69", "sample_id": "d3f8a959cbcb934d7e6d1f6b", "file": "20_seconds/ARCTIC/d3f8a959cbcb934d7e6d1f6b_MANO_GT_20.00s.mp4", "recording_id": "s07/capsulemachine_use_02", "source_first_frame": "3", "source_last_frame": "602", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 35/600 output frames; may include genuine out-of-view frames", "id": "92906ad888eef0bf14cecb69", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/d3f8a959cbcb934d7e6d1f6b_MANO_GT_20.00s.mp4", "bytes": 5009384, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.03807497667376426, "p90": 0.09517469375902694, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 7.491441999533397, "p90": 20.241045270618137, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.06736354322627507, "p90": 0.23937979646417323, "valid_pairs": 563, "valid_duration_s": 18.766666666666666, "coverage": 0.9398998330550917 }, "cmh_m_s": { "median": 0.037350912660047836, "p90": 0.10226940229660277, "valid_pairs": 563, "valid_duration_s": 18.766666666666666, "coverage": 0.9398998330550917 }, "rhcm_continuous_estimate_m_s": { "median": 0.07339990068466462, "p90": 0.1918029160606009, "valid_pairs": 563, "valid_duration_s": 18.766666666666666, "coverage": 0.9398998330550917 }, "rhcm_discrete_m_s": { "median": 0.07314205091559989, "p90": 0.19180451836803344, "valid_pairs": 563, "valid_duration_s": 18.766666666666666, "coverage": 0.9398998330550917 }, "wms": { "median": 0.7242681814187137, "p90": 0.9222510264593701, "valid_pairs": 562, "valid_duration_s": 18.733333333333334, "coverage": 0.9382303839732887 }, "rham_deg_s": { "median": 34.48069400982165, "p90": 92.93649528169264, "valid_pairs": 563, "valid_duration_s": 18.766666666666666, "coverage": 0.9398998330550917 }, "speed_spearman": 0.356031726541485, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.09676184522798842, "p90": 0.2664678465113446, "valid_pairs": 290, "valid_duration_s": 9.666666666666668, "coverage": 0.48414023372287146 }, "cmh_m_s": { "median": 0.03179461545219012, "p90": 0.13404315465374422, "valid_pairs": 290, "valid_duration_s": 9.666666666666668, "coverage": 0.48414023372287146 }, "rhcm_continuous_estimate_m_s": { "median": 0.0801482400962682, "p90": 0.26242728473719995, "valid_pairs": 290, "valid_duration_s": 9.666666666666668, "coverage": 0.48414023372287146 }, "rhcm_discrete_m_s": { "median": 0.08015686507017801, "p90": 0.26258215099550103, "valid_pairs": 290, "valid_duration_s": 9.666666666666668, "coverage": 0.48414023372287146 }, "wms": { "median": 0.7062024775947343, "p90": 0.9399025681163378, "valid_pairs": 290, "valid_duration_s": 9.666666666666668, "coverage": 0.48414023372287146 }, "rham_deg_s": { "median": 48.563015833629834, "p90": 131.92570031020944, "valid_pairs": 290, "valid_duration_s": 9.666666666666668, "coverage": 0.48414023372287146 }, "speed_spearman": 0.48696622330578365, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.09517469375902694, "camera_angular_speed_p90_deg_s": 20.241045270618137, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2664678465113446, "hand_camera_speed_p90_m_s": 0.26258215099550103, "hand_world_valid_fraction": 0.7175, "hand_camera_valid_fraction": 0.7175, "cmh_p90_m_s": 0.13404315465374422, "rhcm_p90_m_s": 0.26258215099550103, "rham_p90_deg_s": 131.92570031020944, "wms_p90": 0.9399025681163378, "gt_coverage": 0.9398998330550917, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "9e9dbe61c98e6f7d0c03715bda5057111cb5632cac1efa1b34258348cdc22404", "metadata_sha256": "e9f5deee49577c775b689e95c2c110ac0a9b170dbf720ed1a7b207f9e6f672c2", "source_file": "long20_20260918/local/train/previews/92906ad888eef0bf14cecb69/reference.npz", "window_s": [ 0.06666666666666667, 20.033333333333335 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.579669976578771e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "55756af658b757525115d4b42e9eadbd1c09df1680ea6392cef55f647abe19b5", "mano_sha256": "5516f597877e26d9eb7aa6957bbdca8316d73c69fec7ea31310fba9819aa3f10", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "94e1f73b2f2e90beae1f55e4", "sample_id": "8239ef01521b3338efbf3f91", "file": "20_seconds/ARCTIC/8239ef01521b3338efbf3f91_MANO_GT_20.00s.mp4", "recording_id": "s05/laptop_use_01", "source_first_frame": "166", "source_last_frame": "765", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 42/600 output frames; may include genuine out-of-view frames", "id": "94e1f73b2f2e90beae1f55e4", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/8239ef01521b3338efbf3f91_MANO_GT_20.00s.mp4", "bytes": 5694753, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.09800134574619401, "p90": 0.19253889770867372, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 17.474591013141936, "p90": 42.4876414067414, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07023753072294109, "p90": 0.28486451792057094, "valid_pairs": 508, "valid_duration_s": 16.933333333333334, "coverage": 0.8480801335559266 }, "cmh_m_s": { "median": 0.09637069457988534, "p90": 0.24211983390684894, "valid_pairs": 508, "valid_duration_s": 16.933333333333334, "coverage": 0.8480801335559266 }, "rhcm_continuous_estimate_m_s": { "median": 0.16238015656679833, "p90": 0.36745367437247556, "valid_pairs": 508, "valid_duration_s": 16.933333333333334, "coverage": 0.8480801335559266 }, "rhcm_discrete_m_s": { "median": 0.1617436532318528, "p90": 0.36623214080825583, "valid_pairs": 508, "valid_duration_s": 16.933333333333334, "coverage": 0.8480801335559266 }, "wms": { "median": 0.8403353185769631, "p90": 0.9751224470830292, "valid_pairs": 508, "valid_duration_s": 16.933333333333334, "coverage": 0.8480801335559266 }, "rham_deg_s": { "median": 42.03174947369255, "p90": 159.85547634418495, "valid_pairs": 508, "valid_duration_s": 16.933333333333334, "coverage": 0.8480801335559266 }, "speed_spearman": -0.09987837615553866, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.17993250247402123, "p90": 0.6569951889384851, "valid_pairs": 557, "valid_duration_s": 18.566666666666666, "coverage": 0.9298831385642738 }, "cmh_m_s": { "median": 0.10255660747825686, "p90": 0.27357155752660367, "valid_pairs": 557, "valid_duration_s": 18.566666666666666, "coverage": 0.9298831385642738 }, "rhcm_continuous_estimate_m_s": { "median": 0.23940451861258832, "p90": 0.6473256661871135, "valid_pairs": 557, "valid_duration_s": 18.566666666666666, "coverage": 0.9298831385642738 }, "rhcm_discrete_m_s": { "median": 0.2393120585059706, "p90": 0.6467864056505086, "valid_pairs": 557, "valid_duration_s": 18.566666666666666, "coverage": 0.9298831385642738 }, "wms": { "median": 0.7795376164587563, "p90": 0.9611229848546466, "valid_pairs": 557, "valid_duration_s": 18.566666666666666, "coverage": 0.9298831385642738 }, "rham_deg_s": { "median": 64.91061947053062, "p90": 171.3245679320316, "valid_pairs": 557, "valid_duration_s": 18.566666666666666, "coverage": 0.9298831385642738 }, "speed_spearman": -0.021753559103258883, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.19253889770867372, "camera_angular_speed_p90_deg_s": 42.4876414067414, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6569951889384851, "hand_camera_speed_p90_m_s": 0.6467864056505086, "hand_world_valid_fraction": 0.89, "hand_camera_valid_fraction": 0.89, "cmh_p90_m_s": 0.27357155752660367, "rhcm_p90_m_s": 0.6467864056505086, "rham_p90_deg_s": 171.3245679320316, "wms_p90": 0.9751224470830292, "gt_coverage": 0.9298831385642738, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "6a7e870dbbb4e6a5785ed9817042f9cb527b75c9b5debc93bfe37def734e7d92", "metadata_sha256": "324b3d92c4f1633fc67065cfdf563eb63f0eb2afeba419f2e8f79c1d86ec8f34", "source_file": "long20_20260918/local/benchmark/previews/94e1f73b2f2e90beae1f55e4/reference.npz", "window_s": [ 5.5, 25.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.577566798522237e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "1b7af68a1c05b1e408b55a6b4f9166bb4a65ee0b29eba08a50b6c5b7bcb6a109", "mano_sha256": "61bc7e050b75c86818fd441bd9012a89a7518dee7c67c70a062e8bd530f196b3", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "98860daede1a079134d8f1cd", "sample_id": "6796c64eb4bd240922891917", "file": "20_seconds/ARCTIC/6796c64eb4bd240922891917_MANO_GT_20.00s.mp4", "recording_id": "s05/espressomachine_use_01", "source_first_frame": "67", "source_last_frame": "666", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 66/600 output frames; may include genuine out-of-view frames", "id": "98860daede1a079134d8f1cd", "aliases": [ "20_seconds/ARCTIC/d4a02910b626511793705685_MANO_GT_20.00s.mp4" ], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/6796c64eb4bd240922891917_MANO_GT_20.00s.mp4", "bytes": 5426660, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08347445926327017, "p90": 0.21045795520429816, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 15.824819867573442, "p90": 53.75306029078894, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.04021939383991593, "p90": 0.2616793597242877, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.8580968280467445 }, "cmh_m_s": { "median": 0.09969589753412011, "p90": 0.252901861565252, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.8580968280467445 }, "rhcm_continuous_estimate_m_s": { "median": 0.1368928145777868, "p90": 0.31018389559460924, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.8580968280467445 }, "rhcm_discrete_m_s": { "median": 0.13685996795179622, "p90": 0.31023553756900124, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.8580968280467445 }, "wms": { "median": 0.8496037619015324, "p90": 0.9636730585567805, "valid_pairs": 513, "valid_duration_s": 17.1, "coverage": 0.8564273789649416 }, "rham_deg_s": { "median": 30.52886556285005, "p90": 126.93390448164261, "valid_pairs": 514, "valid_duration_s": 17.133333333333333, "coverage": 0.8580968280467445 }, "speed_spearman": 0.026417655335497902, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1679441040308256, "p90": 0.5666678817137699, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.8747913188647747 }, "cmh_m_s": { "median": 0.11025698989641886, "p90": 0.3219779269434207, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.8747913188647747 }, "rhcm_continuous_estimate_m_s": { "median": 0.2071219443952489, "p90": 0.7940530842175441, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.8747913188647747 }, "rhcm_discrete_m_s": { "median": 0.20762713087607482, "p90": 0.7892144407960634, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.8747913188647747 }, "wms": { "median": 0.7637824691993983, "p90": 0.9703957611634058, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8731218697829717 }, "rham_deg_s": { "median": 44.546148311264744, "p90": 137.1225759729296, "valid_pairs": 524, "valid_duration_s": 17.46666666666667, "coverage": 0.8747913188647747 }, "speed_spearman": 0.26845182666063366, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.21045795520429816, "camera_angular_speed_p90_deg_s": 53.75306029078894, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5666678817137699, "hand_camera_speed_p90_m_s": 0.7892144407960634, "hand_world_valid_fraction": 0.8683333333333333, "hand_camera_valid_fraction": 0.8683333333333333, "cmh_p90_m_s": 0.3219779269434207, "rhcm_p90_m_s": 0.7892144407960634, "rham_p90_deg_s": 137.1225759729296, "wms_p90": 0.9703957611634058, "gt_coverage": 0.8747913188647747, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "5fa6ad25b31489476a1987f821b0821fd66e37fc21229fd48be21d3c72ce63ff", "metadata_sha256": "e7d3f285b74298553e476b8c10890bc15277b9a0713f2e60f2c5b78e2a13fca5", "source_file": "long20_20260918/local/benchmark/previews/98860daede1a079134d8f1cd/reference.npz", "window_s": [ 2.2, 22.166666666666668 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.585973723508303e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2b33da7ba3e3e756c8b7e3b4b47723bcbad3fcbf02f120f815b1316bad49ca2f", "mano_sha256": "a3bb6983fc6d40683fe85066b8ea3a30e57d54b51fac8df487e091e02a0ac7a5", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "98acdb86226c96983c63a541", "sample_id": "9c47a189d3001e1e18ea3add", "file": "20_seconds/ARCTIC/9c47a189d3001e1e18ea3add_MANO_GT_20.00s.mp4", "recording_id": "s05/microwave_use_02", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 34/600 output frames; may include genuine out-of-view frames", "id": "98acdb86226c96983c63a541", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/9c47a189d3001e1e18ea3add_MANO_GT_20.00s.mp4", "bytes": 5518415, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.11286646540839093, "p90": 0.2571866664741035, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 23.53804355999449, "p90": 45.108985396923984, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.12792599652212208, "p90": 0.6024692556088239, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "cmh_m_s": { "median": 0.15493013043716605, "p90": 0.3227939136424961, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "rhcm_continuous_estimate_m_s": { "median": 0.22088776641542016, "p90": 0.552786428612029, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "rhcm_discrete_m_s": { "median": 0.2207471744060695, "p90": 0.5541077128797811, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "wms": { "median": 0.6934898889079751, "p90": 0.9356394107383129, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "rham_deg_s": { "median": 55.161896500953496, "p90": 137.54871420889003, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741235 }, "speed_spearman": 0.24609996171368725, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.06805274081844032, "p90": 0.7646129559937727, "valid_pairs": 468, "valid_duration_s": 15.599999999999998, "coverage": 0.7813021702838063 }, "cmh_m_s": { "median": 0.14101828749111134, "p90": 0.3020758680298026, "valid_pairs": 468, "valid_duration_s": 15.599999999999998, "coverage": 0.7813021702838063 }, "rhcm_continuous_estimate_m_s": { "median": 0.18556401730068808, "p90": 0.757469149796272, "valid_pairs": 468, "valid_duration_s": 15.599999999999998, "coverage": 0.7813021702838063 }, "rhcm_discrete_m_s": { "median": 0.18546418814079452, "p90": 0.7570837811912412, "valid_pairs": 468, "valid_duration_s": 15.599999999999998, "coverage": 0.7813021702838063 }, "wms": { "median": 0.8029731017676562, "p90": 0.9694219818452487, "valid_pairs": 468, "valid_duration_s": 15.599999999999998, "coverage": 0.7813021702838063 }, "rham_deg_s": { "median": 46.67274712963324, "p90": 181.28087042798646, "valid_pairs": 468, "valid_duration_s": 15.599999999999998, "coverage": 0.7813021702838063 }, "speed_spearman": 0.32173415343219425, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.2571866664741035, "camera_angular_speed_p90_deg_s": 45.108985396923984, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.7646129559937727, "hand_camera_speed_p90_m_s": 0.7570837811912412, "hand_world_valid_fraction": 0.82, "hand_camera_valid_fraction": 0.82, "cmh_p90_m_s": 0.3227939136424961, "rhcm_p90_m_s": 0.7570837811912412, "rham_p90_deg_s": 181.28087042798646, "wms_p90": 0.9694219818452487, "gt_coverage": 0.8464106844741235, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "7258f881fb4e11343bb27b316b28c137ed20bf1463492477a71fb41d7a9a9137", "metadata_sha256": "5d5c8ae8c6538b764bd97b8401b8ab7df79bbee96a1490046315345f6e57bda7", "source_file": "long20_20260918/local/benchmark/previews/98acdb86226c96983c63a541/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.585973723508303e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e3770590426347e781efb88d28290856fbc3143dc8fe9a01e22f761bc71375c9", "mano_sha256": "38a696dae1a42e7dcdfe6838d5a9a98f9f6a14ff314bf5b7fbc390d219860cb6", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "9a6625ba474dd9892abfcbfb", "sample_id": "9061e425557a2995a53506a9", "file": "20_seconds/ARCTIC/9061e425557a2995a53506a9_MANO_GT_20.00s.mp4", "recording_id": "s05/box_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 20/600 output frames; may include genuine out-of-view frames", "id": "9a6625ba474dd9892abfcbfb", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/9061e425557a2995a53506a9_MANO_GT_20.00s.mp4", "bytes": 5982050, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08107529361072469, "p90": 0.16134050906918107, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 17.043305025338285, "p90": 37.563244842590954, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.12480093575163327, "p90": 0.40091990200388783, "valid_pairs": 569, "valid_duration_s": 18.96666666666666, "coverage": 0.9499165275459097 }, "cmh_m_s": { "median": 0.10568794216716726, "p90": 0.23237086741286778, "valid_pairs": 569, "valid_duration_s": 18.96666666666666, "coverage": 0.9499165275459097 }, "rhcm_continuous_estimate_m_s": { "median": 0.17199222540686612, "p90": 0.4489174770259673, "valid_pairs": 569, "valid_duration_s": 18.96666666666666, "coverage": 0.9499165275459097 }, "rhcm_discrete_m_s": { "median": 0.17127183305896473, "p90": 0.4486279784166715, "valid_pairs": 569, "valid_duration_s": 18.96666666666666, "coverage": 0.9499165275459097 }, "wms": { "median": 0.7782335702478785, "p90": 0.9470312481004894, "valid_pairs": 569, "valid_duration_s": 18.96666666666666, "coverage": 0.9499165275459097 }, "rham_deg_s": { "median": 44.827356562950655, "p90": 153.03337417482317, "valid_pairs": 569, "valid_duration_s": 18.96666666666666, "coverage": 0.9499165275459097 }, "speed_spearman": 0.09077289562925607, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.22118594541709258, "p90": 0.511292914255008, "valid_pairs": 488, "valid_duration_s": 16.266666666666666, "coverage": 0.8146911519198665 }, "cmh_m_s": { "median": 0.0973063416869294, "p90": 0.22463160645761365, "valid_pairs": 488, "valid_duration_s": 16.266666666666666, "coverage": 0.8146911519198665 }, "rhcm_continuous_estimate_m_s": { "median": 0.20100401741203128, "p90": 0.527685512086094, "valid_pairs": 488, "valid_duration_s": 16.266666666666666, "coverage": 0.8146911519198665 }, "rhcm_discrete_m_s": { "median": 0.2009804224435034, "p90": 0.5264441498817017, "valid_pairs": 488, "valid_duration_s": 16.266666666666666, "coverage": 0.8146911519198665 }, "wms": { "median": 0.739991698500962, "p90": 0.9392573129975162, "valid_pairs": 488, "valid_duration_s": 16.266666666666666, "coverage": 0.8146911519198665 }, "rham_deg_s": { "median": 52.140599094522386, "p90": 153.48015373203802, "valid_pairs": 488, "valid_duration_s": 16.266666666666666, "coverage": 0.8146911519198665 }, "speed_spearman": 0.2069126670894737, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.16134050906918107, "camera_angular_speed_p90_deg_s": 37.563244842590954, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.511292914255008, "hand_camera_speed_p90_m_s": 0.5264441498817017, "hand_world_valid_fraction": 0.8875, "hand_camera_valid_fraction": 0.8875, "cmh_p90_m_s": 0.23237086741286778, "rhcm_p90_m_s": 0.5264441498817017, "rham_p90_deg_s": 153.48015373203802, "wms_p90": 0.9470312481004894, "gt_coverage": 0.9499165275459097, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "79796d28066edc1b2cf2b5d6cf06c79369711e72d2259d1ed7eb72ee993febf6", "metadata_sha256": "4fe07b5e889741f1bb30d2a5587843a9b3ca548e9d1128ad1238bdb9aa191f6f", "source_file": "long20_20260918/local/benchmark/previews/9a6625ba474dd9892abfcbfb/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.660953740672398e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2edf2748f54d024429d2a2e6e4ce0fde5fc8fcd57aedd0f8a193146232ad61a9", "mano_sha256": "e735580063f5d5d65e5dde3579497136a5374eb951e166c3fe29a1afa1614360", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "False", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "9e639e29705644e13960f012", "sample_id": "3767e67e17e1f71c2a83447d", "file": "20_seconds/ARCTIC/3767e67e17e1f71c2a83447d_MANO_GT_20.00s.mp4", "recording_id": "s05/box_use_02", "source_first_frame": "52", "source_last_frame": "651", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 74/600 output frames; may include genuine out-of-view frames", "id": "9e639e29705644e13960f012", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/3767e67e17e1f71c2a83447d_MANO_GT_20.00s.mp4", "bytes": 5839012, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.07735797387271116, "p90": 0.14123901754394616, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 16.31683356642779, "p90": 33.77896128658384, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.1721972160078661, "p90": 0.43012759768000425, "valid_pairs": 525, "valid_duration_s": 17.5, "coverage": 0.8764607679465776 }, "cmh_m_s": { "median": 0.11412543732128526, "p90": 0.24063639911275042, "valid_pairs": 525, "valid_duration_s": 17.5, "coverage": 0.8764607679465776 }, "rhcm_continuous_estimate_m_s": { "median": 0.20872764779474426, "p90": 0.5371018157507609, "valid_pairs": 525, "valid_duration_s": 17.5, "coverage": 0.8764607679465776 }, "rhcm_discrete_m_s": { "median": 0.20948679226016004, "p90": 0.5371965111321954, "valid_pairs": 525, "valid_duration_s": 17.5, "coverage": 0.8764607679465776 }, "wms": { "median": 0.772241778751867, "p90": 0.9619255725883045, "valid_pairs": 525, "valid_duration_s": 17.5, "coverage": 0.8764607679465776 }, "rham_deg_s": { "median": 57.5064871500685, "p90": 196.12893700359095, "valid_pairs": 525, "valid_duration_s": 17.5, "coverage": 0.8764607679465776 }, "speed_spearman": 0.08577971646673938, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.15657459617395347, "p90": 0.42668669498022116, "valid_pairs": 432, "valid_duration_s": 14.399999999999999, "coverage": 0.7212020033388981 }, "cmh_m_s": { "median": 0.10497369608995555, "p90": 0.1992116446051885, "valid_pairs": 432, "valid_duration_s": 14.399999999999999, "coverage": 0.7212020033388981 }, "rhcm_continuous_estimate_m_s": { "median": 0.2099515218206474, "p90": 0.5175810059725447, "valid_pairs": 432, "valid_duration_s": 14.399999999999999, "coverage": 0.7212020033388981 }, "rhcm_discrete_m_s": { "median": 0.2106394649491734, "p90": 0.5176799153278908, "valid_pairs": 432, "valid_duration_s": 14.399999999999999, "coverage": 0.7212020033388981 }, "wms": { "median": 0.7752038353143202, "p90": 0.9450432605832774, "valid_pairs": 432, "valid_duration_s": 14.399999999999999, "coverage": 0.7212020033388981 }, "rham_deg_s": { "median": 48.8941395648567, "p90": 163.5214212618278, "valid_pairs": 432, "valid_duration_s": 14.399999999999999, "coverage": 0.7212020033388981 }, "speed_spearman": 0.12462528436711684, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.14123901754394616, "camera_angular_speed_p90_deg_s": 33.77896128658384, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.43012759768000425, "hand_camera_speed_p90_m_s": 0.5371965111321954, "hand_world_valid_fraction": 0.8041666666666667, "hand_camera_valid_fraction": 0.8041666666666667, "cmh_p90_m_s": 0.24063639911275042, "rhcm_p90_m_s": 0.5371965111321954, "rham_p90_deg_s": 196.12893700359095, "wms_p90": 0.9619255725883045, "gt_coverage": 0.8764607679465776, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "27317e41d305369c7de4707c24a61564d6479fa27846f5d9bf72bbb1859f0773", "metadata_sha256": "ec4adaa9755d9fee7f72b4d85ee636e55eb6bc2a879b522b477aa557e20b0942", "source_file": "long20_20260918/local/benchmark/previews/9e639e29705644e13960f012/reference.npz", "window_s": [ 1.7, 21.666666666666668 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.742874840267547e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6d0826c29b4c3c007ec5d3232bb2224b59a0de4119398834b68c42b155b3307b", "mano_sha256": "4618f9b45b4d858c2e93bca38c64cf4f330b1e634b4dffcd32772bcde31db1af", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "9ebb3a279c1abb84e8a268bd", "sample_id": "a7ffbee43299eef38743c2f2", "file": "20_seconds/ARCTIC/a7ffbee43299eef38743c2f2_MANO_GT_20.00s.mp4", "recording_id": "s02/laptop_use_01", "source_first_frame": "61", "source_last_frame": "660", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 76/600 output frames; may include genuine out-of-view frames", "id": "9ebb3a279c1abb84e8a268bd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/a7ffbee43299eef38743c2f2_MANO_GT_20.00s.mp4", "bytes": 4799538, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.018984133165674908, "p90": 0.05755628850422161, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.4106183243765082, "p90": 7.804456279858567, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.03626567303765937, "p90": 0.29689872904953735, "valid_pairs": 523, "valid_duration_s": 17.43333333333333, "coverage": 0.8731218697829713 }, "cmh_m_s": { "median": 0.027990662481463518, "p90": 0.05618569197641243, "valid_pairs": 523, "valid_duration_s": 17.43333333333333, "coverage": 0.8731218697829713 }, "rhcm_continuous_estimate_m_s": { "median": 0.038242728733841325, "p90": 0.28694715174053104, "valid_pairs": 523, "valid_duration_s": 17.43333333333333, "coverage": 0.8731218697829713 }, "rhcm_discrete_m_s": { "median": 0.03825373880106619, "p90": 0.28693163579023045, "valid_pairs": 523, "valid_duration_s": 17.43333333333333, "coverage": 0.8731218697829713 }, "wms": { "median": 0.652921932260117, "p90": 0.9341081906120676, "valid_pairs": 516, "valid_duration_s": 17.2, "coverage": 0.8614357262103505 }, "rham_deg_s": { "median": 14.369275274358117, "p90": 79.97106363356814, "valid_pairs": 523, "valid_duration_s": 17.43333333333333, "coverage": 0.8731218697829713 }, "speed_spearman": 0.4626647749206418, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.19362644710448743, "p90": 0.5608129714040866, "valid_pairs": 395, "valid_duration_s": 13.16666666666667, "coverage": 0.6594323873121871 }, "cmh_m_s": { "median": 0.027494509390151214, "p90": 0.048580156242822005, "valid_pairs": 395, "valid_duration_s": 13.16666666666667, "coverage": 0.6594323873121871 }, "rhcm_continuous_estimate_m_s": { "median": 0.19216231039770718, "p90": 0.5600343008980039, "valid_pairs": 395, "valid_duration_s": 13.16666666666667, "coverage": 0.6594323873121871 }, "rhcm_discrete_m_s": { "median": 0.19216253339552136, "p90": 0.5600176742730021, "valid_pairs": 395, "valid_duration_s": 13.16666666666667, "coverage": 0.6594323873121871 }, "wms": { "median": 0.8842968361528386, "p90": 0.9796724907931037, "valid_pairs": 395, "valid_duration_s": 13.16666666666667, "coverage": 0.6594323873121871 }, "rham_deg_s": { "median": 50.48969759955378, "p90": 131.46688476412663, "valid_pairs": 395, "valid_duration_s": 13.16666666666667, "coverage": 0.6594323873121871 }, "speed_spearman": -0.03215902519378713, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05755628850422161, "camera_angular_speed_p90_deg_s": 7.804456279858567, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5608129714040866, "hand_camera_speed_p90_m_s": 0.5600176742730021, "hand_world_valid_fraction": 0.7708333333333334, "hand_camera_valid_fraction": 0.7708333333333334, "cmh_p90_m_s": 0.05618569197641243, "rhcm_p90_m_s": 0.5600176742730021, "rham_p90_deg_s": 131.46688476412663, "wms_p90": 0.9796724907931037, "gt_coverage": 0.8731218697829713, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "b9ff7e188308d4f93bc6d3bd0d29db09192cd4e1a6c6f0eee2324af044b9cfad", "metadata_sha256": "a6527b872d4ff246a580b984114c81dca64b1535f2be7f4287765538c68c65aa", "source_file": "long20_20260918/local/train/previews/9ebb3a279c1abb84e8a268bd/reference.npz", "window_s": [ 1.9666666666666666, 21.933333333333334 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.484050470211362e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "751cd4d4f08f2566965e64f04f456771ceee97db99cfc190dfeeeafdc480bb1b", "mano_sha256": "ac776bb1327b3b763bbbb01b152e6cf875c4a2b96a6c1e8f11ff4204dc99bb67", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a493dc2c58c7339dbed67967", "sample_id": "f94bf0be96cfaacf87d96507", "file": "20_seconds/ARCTIC/f94bf0be96cfaacf87d96507_MANO_GT_20.00s.mp4", "recording_id": "s07/ketchup_use_02", "source_first_frame": "3", "source_last_frame": "602", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 58/600 output frames; may include genuine out-of-view frames", "id": "a493dc2c58c7339dbed67967", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/f94bf0be96cfaacf87d96507_MANO_GT_20.00s.mp4", "bytes": 5216493, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0733468011168549, "p90": 0.16161454609327508, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 13.158302262599303, "p90": 30.16897891342589, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.19299400628025223, "p90": 0.4501051437170028, "valid_pairs": 453, "valid_duration_s": 15.100000000000001, "coverage": 0.7562604340567612 }, "cmh_m_s": { "median": 0.07882752521832705, "p90": 0.18667247171357718, "valid_pairs": 453, "valid_duration_s": 15.100000000000001, "coverage": 0.7562604340567612 }, "rhcm_continuous_estimate_m_s": { "median": 0.1825195152789728, "p90": 0.4197102162983363, "valid_pairs": 453, "valid_duration_s": 15.100000000000001, "coverage": 0.7562604340567612 }, "rhcm_discrete_m_s": { "median": 0.18206866539691455, "p90": 0.4190772214114916, "valid_pairs": 453, "valid_duration_s": 15.100000000000001, "coverage": 0.7562604340567612 }, "wms": { "median": 0.6996452125102108, "p90": 0.9393538541489506, "valid_pairs": 452, "valid_duration_s": 15.066666666666668, "coverage": 0.7545909849749582 }, "rham_deg_s": { "median": 57.38242826455223, "p90": 158.32442718775133, "valid_pairs": 453, "valid_duration_s": 15.100000000000001, "coverage": 0.7562604340567612 }, "speed_spearman": 0.3027544205833732, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.18223001841555284, "p90": 0.46407454487737015, "valid_pairs": 331, "valid_duration_s": 11.03333333333333, "coverage": 0.5525876460767944 }, "cmh_m_s": { "median": 0.061758980951045116, "p90": 0.15266201952449032, "valid_pairs": 331, "valid_duration_s": 11.03333333333333, "coverage": 0.5525876460767944 }, "rhcm_continuous_estimate_m_s": { "median": 0.1644527624512571, "p90": 0.42348169479153447, "valid_pairs": 331, "valid_duration_s": 11.03333333333333, "coverage": 0.5525876460767944 }, "rhcm_discrete_m_s": { "median": 0.1644173102495702, "p90": 0.4227632321420092, "valid_pairs": 331, "valid_duration_s": 11.03333333333333, "coverage": 0.5525876460767944 }, "wms": { "median": 0.6712692097471802, "p90": 0.9576682151524223, "valid_pairs": 331, "valid_duration_s": 11.03333333333333, "coverage": 0.5525876460767944 }, "rham_deg_s": { "median": 65.35641947467946, "p90": 166.45118023983298, "valid_pairs": 331, "valid_duration_s": 11.03333333333333, "coverage": 0.5525876460767944 }, "speed_spearman": 0.48275000579080946, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.16161454609327508, "camera_angular_speed_p90_deg_s": 30.16897891342589, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.46407454487737015, "hand_camera_speed_p90_m_s": 0.4227632321420092, "hand_world_valid_fraction": 0.6683333333333333, "hand_camera_valid_fraction": 0.6683333333333333, "cmh_p90_m_s": 0.18667247171357718, "rhcm_p90_m_s": 0.4227632321420092, "rham_p90_deg_s": 166.45118023983298, "wms_p90": 0.9576682151524223, "gt_coverage": 0.7562604340567612, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "d9ef3d04400a3038770ac8846ce63113e7530a32ee49500778b279626252c19c", "metadata_sha256": "442c0e1956021aa5cba957fcbd0fcefecf33067b63c6d79c1429a6215fa7f659", "source_file": "long20_20260918/local/train/previews/a493dc2c58c7339dbed67967/reference.npz", "window_s": [ 0.06666666666666667, 20.033333333333335 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.75098925995612e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7c622ec7db65f9bd8c62778c869631364e8f4bfbacff9f7ed4f628f03e3dc0ef", "mano_sha256": "321f1ae9e5c24fcdcf1ca8cac395f30bc25b89c6499d5f2fde4bdc5b9efe6ad4", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "a64528dfaf4151943a8205bd", "sample_id": "636b8a631c1ab1759d60dae3", "file": "20_seconds/ARCTIC/636b8a631c1ab1759d60dae3_MANO_GT_20.00s.mp4", "recording_id": "s01/phone_grab_01", "source_first_frame": "130", "source_last_frame": "729", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 39/600 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 7 frames", "id": "a64528dfaf4151943a8205bd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/636b8a631c1ab1759d60dae3_MANO_GT_20.00s.mp4", "bytes": 5002972, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.022956808564166473, "p90": 0.04530454915155111, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 4.4872588936914015, "p90": 10.679416379989558, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.08726380277333406, "p90": 0.44244027934741015, "valid_pairs": 401, "valid_duration_s": 13.366666666666667, "coverage": 0.6694490818030051 }, "cmh_m_s": { "median": 0.030342467891783905, "p90": 0.0676788140184555, "valid_pairs": 401, "valid_duration_s": 13.366666666666667, "coverage": 0.6694490818030051 }, "rhcm_continuous_estimate_m_s": { "median": 0.09467083182073137, "p90": 0.46887681022024547, "valid_pairs": 401, "valid_duration_s": 13.366666666666667, "coverage": 0.6694490818030051 }, "rhcm_discrete_m_s": { "median": 0.09467264537120268, "p90": 0.4689050388062856, "valid_pairs": 401, "valid_duration_s": 13.366666666666667, "coverage": 0.6694490818030051 }, "wms": { "median": 0.8281194345639732, "p90": 0.9774121243932267, "valid_pairs": 391, "valid_duration_s": 13.033333333333339, "coverage": 0.6527545909849752 }, "rham_deg_s": { "median": 39.41153754742899, "p90": 109.82224174931348, "valid_pairs": 401, "valid_duration_s": 13.366666666666667, "coverage": 0.6694490818030051 }, "speed_spearman": 0.3482722298730786, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08943300450112318, "p90": 0.2428083994500826, "valid_pairs": 556, "valid_duration_s": 18.53333333333333, "coverage": 0.9282136894824707 }, "cmh_m_s": { "median": 0.03211198707521527, "p90": 0.06229716120647652, "valid_pairs": 556, "valid_duration_s": 18.53333333333333, "coverage": 0.9282136894824707 }, "rhcm_continuous_estimate_m_s": { "median": 0.08908311356297835, "p90": 0.249003296900912, "valid_pairs": 556, "valid_duration_s": 18.53333333333333, "coverage": 0.9282136894824707 }, "rhcm_discrete_m_s": { "median": 0.08909193693651661, "p90": 0.24896747735852442, "valid_pairs": 556, "valid_duration_s": 18.53333333333333, "coverage": 0.9282136894824707 }, "wms": { "median": 0.7790648069132211, "p90": 0.9491873726593506, "valid_pairs": 554, "valid_duration_s": 18.466666666666665, "coverage": 0.9248747913188647 }, "rham_deg_s": { "median": 54.95270777069173, "p90": 157.67369209595583, "valid_pairs": 556, "valid_duration_s": 18.53333333333333, "coverage": 0.9282136894824707 }, "speed_spearman": 0.37274481254435526, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04530454915155111, "camera_angular_speed_p90_deg_s": 10.679416379989558, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.44244027934741015, "hand_camera_speed_p90_m_s": 0.4689050388062856, "hand_world_valid_fraction": 0.8033333333333333, "hand_camera_valid_fraction": 0.8033333333333333, "cmh_p90_m_s": 0.0676788140184555, "rhcm_p90_m_s": 0.4689050388062856, "rham_p90_deg_s": 157.67369209595583, "wms_p90": 0.9774121243932267, "gt_coverage": 0.9282136894824707, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "f651a4f81de638406377d5cbe58fc9cb3bb3e9157e54b49d623ff3c9514d4ecc", "metadata_sha256": "faf325ca459280d5fa720cfb3193d4865c72f62653119caef1718e1accac1e3a", "source_file": "long20_20260918/local/train/previews/a64528dfaf4151943a8205bd/reference.npz", "window_s": [ 4.3, 24.266666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.484050470211362e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "864cbff977c83ebfcec7bb0dacf51ef11fb53d65b63543c1a7727e8540f13b03", "mano_sha256": "705a812f56f89303fb6f1146b02ff261684e0b5533ad7ee8db06cd4f491d4093", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "a78ed5908078934c8cfe1986", "sample_id": "08f4b5808b449f8f68a48b31", "file": "20_seconds/ARCTIC/08f4b5808b449f8f68a48b31_MANO_GT_20.00s.mp4", "recording_id": "s05/mixer_use_02", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 22/600 output frames; may include genuine out-of-view frames", "id": "a78ed5908078934c8cfe1986", "aliases": [ "20_seconds/ARCTIC/e608e4881bc7a11fdd41a4cd_MANO_GT_20.00s.mp4" ], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/08f4b5808b449f8f68a48b31_MANO_GT_20.00s.mp4", "bytes": 5041972, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.058832836293105394, "p90": 0.12673893394875274, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 9.661926617042662, "p90": 25.22664232173061, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.07375153807151773, "p90": 0.3361755683667416, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "cmh_m_s": { "median": 0.05832515724751663, "p90": 0.12229821524468142, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rhcm_continuous_estimate_m_s": { "median": 0.0972013856306686, "p90": 0.3178410243755216, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rhcm_discrete_m_s": { "median": 0.09725066894740131, "p90": 0.3177070862537489, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "wms": { "median": 0.7609238202328217, "p90": 0.9446214924462819, "valid_pairs": 575, "valid_duration_s": 19.166666666666664, "coverage": 0.9599332220367278 }, "rham_deg_s": { "median": 34.848250380294054, "p90": 115.62491980436867, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "speed_spearman": 0.3106297188056587, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.11325062729466467, "p90": 0.3491403522581449, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "cmh_m_s": { "median": 0.06143953574316081, "p90": 0.12649006474524357, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rhcm_continuous_estimate_m_s": { "median": 0.1285500597505018, "p90": 0.3379604978866895, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rhcm_discrete_m_s": { "median": 0.12777933742116046, "p90": 0.3379341647218445, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "wms": { "median": 0.8279690742440227, "p90": 0.9527229851789363, "valid_pairs": 576, "valid_duration_s": 19.199999999999996, "coverage": 0.9616026711185307 }, "rham_deg_s": { "median": 41.78254738600014, "p90": 118.91326859097103, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "speed_spearman": -0.020203409633809687, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.12673893394875274, "camera_angular_speed_p90_deg_s": 25.22664232173061, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3491403522581449, "hand_camera_speed_p90_m_s": 0.3379341647218445, "hand_world_valid_fraction": 0.9633333333333334, "hand_camera_valid_fraction": 0.9633333333333334, "cmh_p90_m_s": 0.12649006474524357, "rhcm_p90_m_s": 0.3379341647218445, "rham_p90_deg_s": 118.91326859097103, "wms_p90": 0.9527229851789363, "gt_coverage": 0.963272120200334, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "b07f2145cb751c865fa2f2a614be6b9ad0545674c43d145be93e467928be6a08", "metadata_sha256": "3be4a5a4e87ce51b81bef3d20f7018d8901dea187a392571ef6e350445561e03", "source_file": "long20_20260918/local/benchmark/previews/a78ed5908078934c8cfe1986/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.577566798522237e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a8270dee1f692523c73ab9206b68a3c2d9bbee0ddddcd7766c89fddbe5c88c05", "mano_sha256": "d9ff85c43a007922965577398b61da2395e80a75f875398cc3d666b798d0c22b", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "a7fb00f50606ede97b5bc4ac", "sample_id": "cd800b138e62504e5bf087eb", "file": "20_seconds/ARCTIC/cd800b138e62504e5bf087eb_MANO_GT_20.00s.mp4", "recording_id": "s05/microwave_grab_01", "source_first_frame": "76", "source_last_frame": "675", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 54/600 output frames; may include genuine out-of-view frames", "id": "a7fb00f50606ede97b5bc4ac", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/cd800b138e62504e5bf087eb_MANO_GT_20.00s.mp4", "bytes": 5725740, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.14080958591510195, "p90": 0.26573133984422287, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 17.785983418992984, "p90": 39.49208663996196, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.2682549197227738, "p90": 0.6247221495373808, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "cmh_m_s": { "median": 0.12779690783410383, "p90": 0.3013637171402402, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "rhcm_continuous_estimate_m_s": { "median": 0.24454872401801828, "p90": 0.5840296329738577, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "rhcm_discrete_m_s": { "median": 0.2441312263095321, "p90": 0.5846220702085045, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "wms": { "median": 0.6524644529618988, "p90": 0.9182409288599512, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "rham_deg_s": { "median": 70.85207083225373, "p90": 181.2645286328172, "valid_pairs": 502, "valid_duration_s": 16.733333333333334, "coverage": 0.8380634390651086 }, "speed_spearman": 0.2667367695442442, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.30133611100672253, "p90": 0.6081108238041436, "valid_pairs": 526, "valid_duration_s": 17.53333333333333, "coverage": 0.8781302170283806 }, "cmh_m_s": { "median": 0.17502240084515164, "p90": 0.3455840652709988, "valid_pairs": 526, "valid_duration_s": 17.53333333333333, "coverage": 0.8781302170283806 }, "rhcm_continuous_estimate_m_s": { "median": 0.2539934822832919, "p90": 0.5532509888341917, "valid_pairs": 526, "valid_duration_s": 17.53333333333333, "coverage": 0.8781302170283806 }, "rhcm_discrete_m_s": { "median": 0.25391444797761936, "p90": 0.5526462363841989, "valid_pairs": 526, "valid_duration_s": 17.53333333333333, "coverage": 0.8781302170283806 }, "wms": { "median": 0.6173577404492017, "p90": 0.8972939271974121, "valid_pairs": 526, "valid_duration_s": 17.53333333333333, "coverage": 0.8781302170283806 }, "rham_deg_s": { "median": 66.77746366172352, "p90": 172.56665774865192, "valid_pairs": 526, "valid_duration_s": 17.53333333333333, "coverage": 0.8781302170283806 }, "speed_spearman": 0.46283067427878793, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.26573133984422287, "camera_angular_speed_p90_deg_s": 39.49208663996196, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6247221495373808, "hand_camera_speed_p90_m_s": 0.5846220702085045, "hand_world_valid_fraction": 0.8616666666666667, "hand_camera_valid_fraction": 0.8616666666666667, "cmh_p90_m_s": 0.3455840652709988, "rhcm_p90_m_s": 0.5846220702085045, "rham_p90_deg_s": 181.2645286328172, "wms_p90": 0.9182409288599512, "gt_coverage": 0.8781302170283806, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "1d1b51a870f501237725920f83daedc8305d79028381db6c09564e44ee2c3a0b", "metadata_sha256": "3ed63156905403dd23c212cf37411cca1521d331d91f33d257a076438e4878bf", "source_file": "long20_20260918/local/benchmark/previews/a7fb00f50606ede97b5bc4ac/reference.npz", "window_s": [ 2.5, 22.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.652682298944613e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "1ee8cd405c67cf323a53eb89a2d8bc75ddffbf5ef74dd0c57c586bfb0dda5977", "mano_sha256": "9b9737a54cf64d617afa0b16e08d3a1c85a2fb3cc827b7c6380a931f6dfff6e9", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "acc1d0da53e7edf163137b72", "sample_id": "3a3a096bed650e16aa6f5b29", "file": "20_seconds/ARCTIC/3a3a096bed650e16aa6f5b29_MANO_GT_20.00s.mp4", "recording_id": "s05/waffleiron_use_02", "source_first_frame": "223", "source_last_frame": "822", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 53/600 output frames; may include genuine out-of-view frames", "id": "acc1d0da53e7edf163137b72", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/3a3a096bed650e16aa6f5b29_MANO_GT_20.00s.mp4", "bytes": 5799749, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.09141463075139902, "p90": 0.22364026004710508, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 18.17665572060117, "p90": 52.635516443002125, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.1445960300068344, "p90": 0.49051111634516215, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.8981636060100168 }, "cmh_m_s": { "median": 0.11377223955583912, "p90": 0.2309265294503447, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.8981636060100168 }, "rhcm_continuous_estimate_m_s": { "median": 0.18820756657608878, "p90": 0.553725871155911, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.8981636060100168 }, "rhcm_discrete_m_s": { "median": 0.1877889789016064, "p90": 0.554403022070958, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.8981636060100168 }, "wms": { "median": 0.729238214048922, "p90": 0.953426398439731, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.8981636060100168 }, "rham_deg_s": { "median": 85.11763389169886, "p90": 212.7602679872186, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.8981636060100168 }, "speed_spearman": 0.29751979596961303, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.10986300046049549, "p90": 0.3102942089398409, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "cmh_m_s": { "median": 0.09374912688142308, "p90": 0.22203494115586975, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "rhcm_continuous_estimate_m_s": { "median": 0.14610483578720698, "p90": 0.3239575140021356, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "rhcm_discrete_m_s": { "median": 0.14513729134530853, "p90": 0.32436931640320904, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "wms": { "median": 0.701127574028373, "p90": 0.9405236533527453, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "rham_deg_s": { "median": 54.32302691695227, "p90": 149.92244458776463, "valid_pairs": 541, "valid_duration_s": 18.033333333333335, "coverage": 0.9031719532554257 }, "speed_spearman": 0.2163968294026748, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.22364026004710508, "camera_angular_speed_p90_deg_s": 52.635516443002125, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.49051111634516215, "hand_camera_speed_p90_m_s": 0.554403022070958, "hand_world_valid_fraction": 0.9016666666666666, "hand_camera_valid_fraction": 0.9016666666666666, "cmh_p90_m_s": 0.2309265294503447, "rhcm_p90_m_s": 0.554403022070958, "rham_p90_deg_s": 212.7602679872186, "wms_p90": 0.953426398439731, "gt_coverage": 0.9031719532554257, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "1c1d22c9361dff1db1c68f553afd8394cd61b2e4dff4cf17ccc593773bfc2447", "metadata_sha256": "4d4265d627bb5f26cff94fa91c662dc92e784e155207beaf36985e106a0846bc", "source_file": "long20_20260918/local/benchmark/previews/acc1d0da53e7edf163137b72/reference.npz", "window_s": [ 7.4, 27.366666666666667 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.611102534756203e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e84571180ba9eee34fdffde7828fea8f4180bf2ad9bf744597caaf8096742e61", "mano_sha256": "a577f673c8c900891f7e54a1b1f924b015cbf22e84a2c2a8c98a1dd233b98150", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "c74fd3c810066f7e3204aae9", "sample_id": "dc2f0f677ab4b3d04f91ad61", "file": "20_seconds/ARCTIC/dc2f0f677ab4b3d04f91ad61_MANO_GT_20.00s.mp4", "recording_id": "s01/microwave_grab_01", "source_first_frame": "132", "source_last_frame": "731", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 30/600 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 4 frames", "id": "c74fd3c810066f7e3204aae9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/dc2f0f677ab4b3d04f91ad61_MANO_GT_20.00s.mp4", "bytes": 5006310, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02596493937372839, "p90": 0.05328406420811543, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 4.732675383442157, "p90": 13.406017319978144, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.096985426876675, "p90": 0.2260524465511376, "valid_pairs": 566, "valid_duration_s": 18.866666666666664, "coverage": 0.9449081803005007 }, "cmh_m_s": { "median": 0.03703971605407114, "p90": 0.08396029990397239, "valid_pairs": 566, "valid_duration_s": 18.866666666666664, "coverage": 0.9449081803005007 }, "rhcm_continuous_estimate_m_s": { "median": 0.09910072741118056, "p90": 0.2362798975454134, "valid_pairs": 566, "valid_duration_s": 18.866666666666664, "coverage": 0.9449081803005007 }, "rhcm_discrete_m_s": { "median": 0.09913899366914727, "p90": 0.23617187063941292, "valid_pairs": 566, "valid_duration_s": 18.866666666666664, "coverage": 0.9449081803005007 }, "wms": { "median": 0.7889023728897919, "p90": 0.9470423931218388, "valid_pairs": 566, "valid_duration_s": 18.866666666666664, "coverage": 0.9449081803005007 }, "rham_deg_s": { "median": 28.200936518239818, "p90": 77.59275679626495, "valid_pairs": 566, "valid_duration_s": 18.866666666666664, "coverage": 0.9449081803005007 }, "speed_spearman": 0.33329805886171, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.10759871528069903, "p90": 0.2770482425564111, "valid_pairs": 568, "valid_duration_s": 18.93333333333333, "coverage": 0.9482470784641067 }, "cmh_m_s": { "median": 0.03859063866304992, "p90": 0.08916576121912129, "valid_pairs": 568, "valid_duration_s": 18.93333333333333, "coverage": 0.9482470784641067 }, "rhcm_continuous_estimate_m_s": { "median": 0.11989221167913375, "p90": 0.2986439282916954, "valid_pairs": 568, "valid_duration_s": 18.93333333333333, "coverage": 0.9482470784641067 }, "rhcm_discrete_m_s": { "median": 0.11989060762840464, "p90": 0.29857360534684185, "valid_pairs": 568, "valid_duration_s": 18.93333333333333, "coverage": 0.9482470784641067 }, "wms": { "median": 0.8219267106697159, "p90": 0.9518050528119042, "valid_pairs": 565, "valid_duration_s": 18.83333333333333, "coverage": 0.9432387312186977 }, "rham_deg_s": { "median": 26.94441873366032, "p90": 67.31364599550231, "valid_pairs": 568, "valid_duration_s": 18.93333333333333, "coverage": 0.9482470784641067 }, "speed_spearman": 0.31490714776192147, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05328406420811543, "camera_angular_speed_p90_deg_s": 13.406017319978144, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.2770482425564111, "hand_camera_speed_p90_m_s": 0.29857360534684185, "hand_world_valid_fraction": 0.9483333333333334, "hand_camera_valid_fraction": 0.9483333333333334, "cmh_p90_m_s": 0.08916576121912129, "rhcm_p90_m_s": 0.29857360534684185, "rham_p90_deg_s": 77.59275679626495, "wms_p90": 0.9518050528119042, "gt_coverage": 0.9482470784641067, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "1d8471a72bfdc40bca7e084d6be9dbbf68d638bb1d09f3cf323ae158fad9a133", "metadata_sha256": "a14ea03f9457af6e2f38fa321448eb64c2d87866865e8bf5f57a1ec76b7f21f9", "source_file": "long20_20260918/local/train/previews/c74fd3c810066f7e3204aae9/reference.npz", "window_s": [ 4.366666666666666, 24.333333333333332 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.685680459230494e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8e19153c1df7d6b0d66d5b660fca51e7bbe30f6adec34c732677b2254fcb289e", "mano_sha256": "e95ac27b106ccd1f016dd56d6e30f21e2a1d108d8bfd5e1d52a7bb189e437df1", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "c870574a11032b099f5c2de3", "sample_id": "8f708ad965a2cc754a5c19ab", "file": "20_seconds/ARCTIC/8f708ad965a2cc754a5c19ab_MANO_GT_20.00s.mp4", "recording_id": "s09/mixer_use_01", "source_first_frame": "50", "source_last_frame": "649", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 50/600 output frames; may include genuine out-of-view frames", "id": "c870574a11032b099f5c2de3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/8f708ad965a2cc754a5c19ab_MANO_GT_20.00s.mp4", "bytes": 4633927, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.026160831957217693, "p90": 0.05417038455817971, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 4.723930366057152, "p90": 8.741397605995088, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.09252206946227266, "p90": 0.333766743040789, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "cmh_m_s": { "median": 0.038068606856452566, "p90": 0.0699494073978725, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "rhcm_continuous_estimate_m_s": { "median": 0.09290748601455699, "p90": 0.348144040608467, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "rhcm_discrete_m_s": { "median": 0.09277486231279013, "p90": 0.348134074900263, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "wms": { "median": 0.8231194744215432, "p90": 0.9675935270195853, "valid_pairs": 545, "valid_duration_s": 18.166666666666664, "coverage": 0.9098497495826376 }, "rham_deg_s": { "median": 37.42079057613693, "p90": 121.53886145472279, "valid_pairs": 549, "valid_duration_s": 18.299999999999997, "coverage": 0.9165275459098496 }, "speed_spearman": -0.07049223840656174, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.10256710922427582, "p90": 0.343854484242562, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008348 }, "cmh_m_s": { "median": 0.04014846169109745, "p90": 0.07592250495382566, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008348 }, "rhcm_continuous_estimate_m_s": { "median": 0.10743981665582153, "p90": 0.3414303624514309, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008348 }, "rhcm_discrete_m_s": { "median": 0.10742909459421467, "p90": 0.3414171147677588, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008348 }, "wms": { "median": 0.7780516543949192, "p90": 0.9614931708271097, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008348 }, "rham_deg_s": { "median": 53.90377664695719, "p90": 160.1103171039634, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008348 }, "speed_spearman": 0.19747080290882632, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05417038455817971, "camera_angular_speed_p90_deg_s": 8.741397605995088, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.343854484242562, "hand_camera_speed_p90_m_s": 0.348134074900263, "hand_world_valid_fraction": 0.9125, "hand_camera_valid_fraction": 0.9125, "cmh_p90_m_s": 0.07592250495382566, "rhcm_p90_m_s": 0.348134074900263, "rham_p90_deg_s": 160.1103171039634, "wms_p90": 0.9675935270195853, "gt_coverage": 0.9165275459098496, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "738f5555db05c4fa234800b099a4f468fdfe2a369a889552e06dba481dc6fbc0", "metadata_sha256": "11395c87acb752f5729954e437c6ebe767e6c1133695e0e5c031fb3a6c8ee4e0", "source_file": "long20_20260918/local/train/previews/c870574a11032b099f5c2de3/reference.npz", "window_s": [ 1.6, 21.566666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.652682298944613e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "cdbede3887c18b00a53429b40db3b0e052ffbced8f8bc9af6050e19ecc8b501b", "mano_sha256": "51fce7b1a2d9b6cd54eb3a4dbfc7d4fb64cc78b57a2edee7247db1bb5ecaad51", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "ccd1de8a3da2d4402c917da8", "sample_id": "d7064e3496649bae38250d11", "file": "20_seconds/ARCTIC/d7064e3496649bae38250d11_MANO_GT_20.00s.mp4", "recording_id": "s02/mixer_grab_01", "source_first_frame": "354", "source_last_frame": "953", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 137/600 output frames; may include genuine out-of-view frames", "id": "ccd1de8a3da2d4402c917da8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/d7064e3496649bae38250d11_MANO_GT_20.00s.mp4", "bytes": 4308935, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.029863805191543927, "p90": 0.05824623583486539, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 5.468275560968003, "p90": 11.834566952170311, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.20665122681566905, "p90": 0.5709908942535108, "valid_pairs": 446, "valid_duration_s": 14.86666666666666, "coverage": 0.7445742904841398 }, "cmh_m_s": { "median": 0.03451146999874519, "p90": 0.060397366057428827, "valid_pairs": 446, "valid_duration_s": 14.86666666666666, "coverage": 0.7445742904841398 }, "rhcm_continuous_estimate_m_s": { "median": 0.18282270893853292, "p90": 0.5665738995724492, "valid_pairs": 446, "valid_duration_s": 14.86666666666666, "coverage": 0.7445742904841398 }, "rhcm_discrete_m_s": { "median": 0.18284550244247838, "p90": 0.5665947325934169, "valid_pairs": 446, "valid_duration_s": 14.86666666666666, "coverage": 0.7445742904841398 }, "wms": { "median": 0.7808013316027902, "p90": 0.973705336873592, "valid_pairs": 446, "valid_duration_s": 14.86666666666666, "coverage": 0.7445742904841398 }, "rham_deg_s": { "median": 33.80810062211802, "p90": 77.86154753708276, "valid_pairs": 446, "valid_duration_s": 14.86666666666666, "coverage": 0.7445742904841398 }, "speed_spearman": 0.08932636684055845, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1837336926205778, "p90": 0.5574319373208786, "valid_pairs": 289, "valid_duration_s": 9.633333333333336, "coverage": 0.4824707846410686 }, "cmh_m_s": { "median": 0.03586321193052902, "p90": 0.06060054538886451, "valid_pairs": 289, "valid_duration_s": 9.633333333333336, "coverage": 0.4824707846410686 }, "rhcm_continuous_estimate_m_s": { "median": 0.16281574253534464, "p90": 0.5433195120382716, "valid_pairs": 289, "valid_duration_s": 9.633333333333336, "coverage": 0.4824707846410686 }, "rhcm_discrete_m_s": { "median": 0.1628646859098905, "p90": 0.5433151531230432, "valid_pairs": 289, "valid_duration_s": 9.633333333333336, "coverage": 0.4824707846410686 }, "wms": { "median": 0.7902126286715337, "p90": 0.9659351012724271, "valid_pairs": 289, "valid_duration_s": 9.633333333333336, "coverage": 0.4824707846410686 }, "rham_deg_s": { "median": 32.63200398180979, "p90": 73.66218126537088, "valid_pairs": 289, "valid_duration_s": 9.633333333333336, "coverage": 0.4824707846410686 }, "speed_spearman": 0.1831349679831365, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.05824623583486539, "camera_angular_speed_p90_deg_s": 11.834566952170311, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5709908942535108, "hand_camera_speed_p90_m_s": 0.5665947325934169, "hand_world_valid_fraction": 0.62, "hand_camera_valid_fraction": 0.62, "cmh_p90_m_s": 0.06060054538886451, "rhcm_p90_m_s": 0.5665947325934169, "rham_p90_deg_s": 77.86154753708276, "wms_p90": 0.973705336873592, "gt_coverage": 0.7445742904841398, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "588eef30fe1fb9ba6cb51ebe3ca1cade992c3d9365ba758afeabb9d8a6e86e0b", "metadata_sha256": "b6636cc97189425f5f55f255eedac4eb59ad8f5034784ded6677df8e6041aecb", "source_file": "long20_20260918/local/train/previews/ccd1de8a3da2d4402c917da8/reference.npz", "window_s": [ 11.733333333333333, 31.7 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.458205583648681e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6e4fe0921b9247b2d516028b8a1fa9b7509c2a62de2b8cfca45a920dafcbbb3b", "mano_sha256": "b4e41c459f49989824a2abd3559e3ff0e2d9f1edc1351cc17b2d62ea7e1b6320", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "中末帧官方双手表面可见；手物接触和场景地面分层分别审核。", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "cd4f1126a9d68eda634627e3", "sample_id": "87c30984b27c87947d89faf9", "file": "20_seconds/ARCTIC/87c30984b27c87947d89faf9_MANO_GT_20.00s.mp4", "recording_id": "s04/ketchup_grab_01", "source_first_frame": "12", "source_last_frame": "611", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 42/600 output frames; may include genuine out-of-view frames", "id": "cd4f1126a9d68eda634627e3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/87c30984b27c87947d89faf9_MANO_GT_20.00s.mp4", "bytes": 4427540, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.046550252461150396, "p90": 0.10026791134811304, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 9.490049663664186, "p90": 25.02351609796841, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.16510906759496408, "p90": 0.588891636091046, "valid_pairs": 149, "valid_duration_s": 4.966666666666667, "coverage": 0.24874791318864772 }, "cmh_m_s": { "median": 0.0637208992413239, "p90": 0.2010183230181566, "valid_pairs": 149, "valid_duration_s": 4.966666666666667, "coverage": 0.24874791318864772 }, "rhcm_continuous_estimate_m_s": { "median": 0.14655546189337915, "p90": 0.6194507126182736, "valid_pairs": 149, "valid_duration_s": 4.966666666666667, "coverage": 0.24874791318864772 }, "rhcm_discrete_m_s": { "median": 0.14612149436149627, "p90": 0.6193673878559783, "valid_pairs": 149, "valid_duration_s": 4.966666666666667, "coverage": 0.24874791318864772 }, "wms": { "median": 0.7242229359163459, "p90": 0.971677125475064, "valid_pairs": 149, "valid_duration_s": 4.966666666666667, "coverage": 0.24874791318864772 }, "rham_deg_s": { "median": 45.8453866539833, "p90": 216.44011639030103, "valid_pairs": 149, "valid_duration_s": 4.966666666666667, "coverage": 0.24874791318864772 }, "speed_spearman": 0.33041538182477775, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.26275526412488764, "p90": 0.6528614564174319, "valid_pairs": 429, "valid_duration_s": 14.300000000000002, "coverage": 0.7161936560934892 }, "cmh_m_s": { "median": 0.07925766675995652, "p90": 0.20160252920460517, "valid_pairs": 429, "valid_duration_s": 14.300000000000002, "coverage": 0.7161936560934892 }, "rhcm_continuous_estimate_m_s": { "median": 0.20087877921116204, "p90": 0.617840643570049, "valid_pairs": 429, "valid_duration_s": 14.300000000000002, "coverage": 0.7161936560934892 }, "rhcm_discrete_m_s": { "median": 0.20088106963436342, "p90": 0.6178327108035272, "valid_pairs": 429, "valid_duration_s": 14.300000000000002, "coverage": 0.7161936560934892 }, "wms": { "median": 0.7102688770149345, "p90": 0.9168301431843378, "valid_pairs": 429, "valid_duration_s": 14.300000000000002, "coverage": 0.7161936560934892 }, "rham_deg_s": { "median": 64.68865038847474, "p90": 277.1570778338088, "valid_pairs": 429, "valid_duration_s": 14.300000000000002, "coverage": 0.7161936560934892 }, "speed_spearman": 0.44936871431111813, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.10026791134811304, "camera_angular_speed_p90_deg_s": 25.02351609796841, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6528614564174319, "hand_camera_speed_p90_m_s": 0.6193673878559783, "hand_world_valid_fraction": 0.4866666666666667, "hand_camera_valid_fraction": 0.4866666666666667, "cmh_p90_m_s": 0.20160252920460517, "rhcm_p90_m_s": 0.6193673878559783, "rham_p90_deg_s": 277.1570778338088, "wms_p90": 0.971677125475064, "gt_coverage": 0.7161936560934892, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "56435422b564c69cea4c00e381445f580e029dbc99c2610d50e29894999a3b8a", "metadata_sha256": "d9e427915f3268d0014cfdd028a0cffff1dee8fdc8bcf25e6490c5e3d5b7516c", "source_file": "long20_20260918/local/train/previews/cd4f1126a9d68eda634627e3/reference.npz", "window_s": [ 0.3333333333333333, 20.3 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.585973723508303e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "caddd6428eb00583bd169704fc65874918771c199993b2cd0b7c52e7b6225901", "mano_sha256": "012c9a3864243add10294cd2b88ec8be433a5768a36fa37ce4232841af5f4d01", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "d36cefc0daa14c91994a7555", "sample_id": "73ed4ff3f8d1aa66bce3a4c7", "file": "20_seconds/ARCTIC/73ed4ff3f8d1aa66bce3a4c7_MANO_GT_20.00s.mp4", "recording_id": "s05/waffleiron_grab_01", "source_first_frame": "25", "source_last_frame": "624", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 80/600 output frames; may include genuine out-of-view frames", "id": "d36cefc0daa14c91994a7555", "aliases": [ "20_seconds/ARCTIC/80071c771607dd8eb049c56b_MANO_GT_20.00s.mp4" ], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/73ed4ff3f8d1aa66bce3a4c7_MANO_GT_20.00s.mp4", "bytes": 5157974, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.09277215441675551, "p90": 0.1921522373583677, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 18.377957745298357, "p90": 53.23637279988563, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.20146761539576338, "p90": 0.4677325387232414, "valid_pairs": 395, "valid_duration_s": 13.166666666666666, "coverage": 0.659432387312187 }, "cmh_m_s": { "median": 0.11029898279956615, "p90": 0.2810008015360938, "valid_pairs": 395, "valid_duration_s": 13.166666666666666, "coverage": 0.659432387312187 }, "rhcm_continuous_estimate_m_s": { "median": 0.21560361143029047, "p90": 0.46198468675639576, "valid_pairs": 395, "valid_duration_s": 13.166666666666666, "coverage": 0.659432387312187 }, "rhcm_discrete_m_s": { "median": 0.21566298228835518, "p90": 0.46060075447785137, "valid_pairs": 395, "valid_duration_s": 13.166666666666666, "coverage": 0.659432387312187 }, "wms": { "median": 0.680881098083084, "p90": 0.9003870979539099, "valid_pairs": 395, "valid_duration_s": 13.166666666666666, "coverage": 0.659432387312187 }, "rham_deg_s": { "median": 69.20754880219572, "p90": 220.19925346326636, "valid_pairs": 395, "valid_duration_s": 13.166666666666666, "coverage": 0.659432387312187 }, "speed_spearman": 0.48412980281514634, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.24744954243268155, "p90": 0.6457417269870669, "valid_pairs": 379, "valid_duration_s": 12.633333333333333, "coverage": 0.6327212020033389 }, "cmh_m_s": { "median": 0.12830714596846493, "p90": 0.36460542329292994, "valid_pairs": 379, "valid_duration_s": 12.633333333333333, "coverage": 0.6327212020033389 }, "rhcm_continuous_estimate_m_s": { "median": 0.2491351057141616, "p90": 0.5216388925255253, "valid_pairs": 379, "valid_duration_s": 12.633333333333333, "coverage": 0.6327212020033389 }, "rhcm_discrete_m_s": { "median": 0.24937617088900993, "p90": 0.5215758523705295, "valid_pairs": 379, "valid_duration_s": 12.633333333333333, "coverage": 0.6327212020033389 }, "wms": { "median": 0.6899944859274683, "p90": 0.9151269747077959, "valid_pairs": 379, "valid_duration_s": 12.633333333333333, "coverage": 0.6327212020033389 }, "rham_deg_s": { "median": 78.39168787835332, "p90": 182.07642728693008, "valid_pairs": 379, "valid_duration_s": 12.633333333333333, "coverage": 0.6327212020033389 }, "speed_spearman": 0.43127850408783613, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1921522373583677, "camera_angular_speed_p90_deg_s": 53.23637279988563, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.6457417269870669, "hand_camera_speed_p90_m_s": 0.5215758523705295, "hand_world_valid_fraction": 0.6541666666666667, "hand_camera_valid_fraction": 0.6541666666666667, "cmh_p90_m_s": 0.36460542329292994, "rhcm_p90_m_s": 0.5215758523705295, "rham_p90_deg_s": 220.19925346326636, "wms_p90": 0.9151269747077959, "gt_coverage": 0.659432387312187, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "aa86401f55a120503134faf74c0ca161678fcd88178fe289c648b8f0692db6cf", "metadata_sha256": "08219bb957e1c060ef40b6921f2115a4e2dff68e93e79bd5e61fa98f7113fbfe", "source_file": "long20_20260918/local/benchmark/previews/d36cefc0daa14c91994a7555/reference.npz", "window_s": [ 0.8, 20.766666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.654751537338637e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "697afc5000c4b67fc46ef73a5c689bd5ca72b3017b387f9231b31d4dac321b7f", "mano_sha256": "7f31d2078bf789c1b5b456c328d890f3476a0e0e82de9adf4e0b7863306d83da", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "d9341a3bf10cba7f58106b82", "sample_id": "9f5a95307f5d90fb3af4256b", "file": "20_seconds/ARCTIC/9f5a95307f5d90fb3af4256b_MANO_GT_20.00s.mp4", "recording_id": "s07/notebook_use_01", "source_first_frame": "2", "source_last_frame": "601", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 46/600 output frames; may include genuine out-of-view frames", "id": "d9341a3bf10cba7f58106b82", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/9f5a95307f5d90fb3af4256b_MANO_GT_20.00s.mp4", "bytes": 5208567, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.04472928789628481, "p90": 0.20015606960009594, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 8.932750740864918, "p90": 40.99989514734991, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.06695346416662001, "p90": 0.24413734497684123, "valid_pairs": 479, "valid_duration_s": 15.966666666666665, "coverage": 0.7996661101836394 }, "cmh_m_s": { "median": 0.039904588352016726, "p90": 0.13909514914214874, "valid_pairs": 479, "valid_duration_s": 15.966666666666665, "coverage": 0.7996661101836394 }, "rhcm_continuous_estimate_m_s": { "median": 0.07597739399455102, "p90": 0.18144491987235284, "valid_pairs": 479, "valid_duration_s": 15.966666666666665, "coverage": 0.7996661101836394 }, "rhcm_discrete_m_s": { "median": 0.07599301145750686, "p90": 0.18148906528842057, "valid_pairs": 479, "valid_duration_s": 15.966666666666665, "coverage": 0.7996661101836394 }, "wms": { "median": 0.7028084996537411, "p90": 0.8952603522191671, "valid_pairs": 479, "valid_duration_s": 15.966666666666665, "coverage": 0.7996661101836394 }, "rham_deg_s": { "median": 29.35985095739655, "p90": 97.72616082495429, "valid_pairs": 479, "valid_duration_s": 15.966666666666665, "coverage": 0.7996661101836394 }, "speed_spearman": 0.399374896270997, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.13055940055299228, "p90": 0.3192170964866532, "valid_pairs": 153, "valid_duration_s": 5.099999999999996, "coverage": 0.2554257095158596 }, "cmh_m_s": { "median": 0.05497290621882928, "p90": 0.1920556287745386, "valid_pairs": 153, "valid_duration_s": 5.099999999999996, "coverage": 0.2554257095158596 }, "rhcm_continuous_estimate_m_s": { "median": 0.11479703854396646, "p90": 0.33718402588898155, "valid_pairs": 153, "valid_duration_s": 5.099999999999996, "coverage": 0.2554257095158596 }, "rhcm_discrete_m_s": { "median": 0.11479973567442933, "p90": 0.3371825809429538, "valid_pairs": 153, "valid_duration_s": 5.099999999999996, "coverage": 0.2554257095158596 }, "wms": { "median": 0.7592796982825774, "p90": 0.9635643159430959, "valid_pairs": 153, "valid_duration_s": 5.099999999999996, "coverage": 0.2554257095158596 }, "rham_deg_s": { "median": 32.797598949257825, "p90": 88.80234027723277, "valid_pairs": 153, "valid_duration_s": 5.099999999999996, "coverage": 0.2554257095158596 }, "speed_spearman": 0.1341957835765885, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.20015606960009594, "camera_angular_speed_p90_deg_s": 40.99989514734991, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3192170964866532, "hand_camera_speed_p90_m_s": 0.3371825809429538, "hand_world_valid_fraction": 0.5341666666666667, "hand_camera_valid_fraction": 0.5341666666666667, "cmh_p90_m_s": 0.1920556287745386, "rhcm_p90_m_s": 0.3371825809429538, "rham_p90_deg_s": 97.72616082495429, "wms_p90": 0.9635643159430959, "gt_coverage": 0.7996661101836394, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "87d54da4f8c6f408d887ce46f90fe04ed78ca721035882b94b45c86a071f1e24", "metadata_sha256": "a0468266f0c2af3eafc29cd0b0e43fdbd216a17d6ff4a57f7e74ad243cf49970", "source_file": "long20_20260918/local/train/previews/d9341a3bf10cba7f58106b82/reference.npz", "window_s": [ 0.03333333333333333, 20.0 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.554367738890935e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "10781839c682d7c0da154a48c50553c52a49a3bd3a4b8fa5f2e0bda210de8344", "mano_sha256": "d91f9e22154d8e3f5a888b65b7cbc7c5e0485226788a55f9a1b4dbfeac85f856", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "da7beb32799140c1b74bb483", "sample_id": "a7c27b495f19bc6ef5f561c9", "file": "20_seconds/ARCTIC/a7c27b495f19bc6ef5f561c9_MANO_GT_20.00s.mp4", "recording_id": "s04/laptop_grab_01", "source_first_frame": "280", "source_last_frame": "879", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 66/600 output frames; may include genuine out-of-view frames", "id": "da7beb32799140c1b74bb483", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/a7c27b495f19bc6ef5f561c9_MANO_GT_20.00s.mp4", "bytes": 4988068, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.047907227533022914, "p90": 0.09289810114490048, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 9.23068118688522, "p90": 18.400267769578114, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.1997739932715735, "p90": 0.5535372041869571, "valid_pairs": 341, "valid_duration_s": 11.366666666666669, "coverage": 0.5692821368948248 }, "cmh_m_s": { "median": 0.06572130000316329, "p90": 0.10910014460316511, "valid_pairs": 341, "valid_duration_s": 11.366666666666669, "coverage": 0.5692821368948248 }, "rhcm_continuous_estimate_m_s": { "median": 0.1916423470091551, "p90": 0.5589110035881436, "valid_pairs": 341, "valid_duration_s": 11.366666666666669, "coverage": 0.5692821368948248 }, "rhcm_discrete_m_s": { "median": 0.19179345066889003, "p90": 0.5588012599910649, "valid_pairs": 341, "valid_duration_s": 11.366666666666669, "coverage": 0.5692821368948248 }, "wms": { "median": 0.8329599267735711, "p90": 0.9739366538798744, "valid_pairs": 340, "valid_duration_s": 11.333333333333334, "coverage": 0.5676126878130217 }, "rham_deg_s": { "median": 60.464341072792905, "p90": 142.91838553139792, "valid_pairs": 341, "valid_duration_s": 11.366666666666669, "coverage": 0.5692821368948248 }, "speed_spearman": 0.21800659143113954, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.16121822302983996, "p90": 0.518484373941807, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010016 }, "cmh_m_s": { "median": 0.06498800713457638, "p90": 0.13554771702040083, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010016 }, "rhcm_continuous_estimate_m_s": { "median": 0.15859520887299416, "p90": 0.4924139106477104, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010016 }, "rhcm_discrete_m_s": { "median": 0.15818472356231295, "p90": 0.49236820341540416, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010016 }, "wms": { "median": 0.7341308042742967, "p90": 0.9243510777673539, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010016 }, "rham_deg_s": { "median": 64.7430703257645, "p90": 153.91361168703025, "valid_pairs": 533, "valid_duration_s": 17.766666666666666, "coverage": 0.8898163606010016 }, "speed_spearman": 0.3081199425403317, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "low", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.09289810114490048, "camera_angular_speed_p90_deg_s": 18.400267769578114, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5535372041869571, "hand_camera_speed_p90_m_s": 0.5588012599910649, "hand_world_valid_fraction": 0.7325, "hand_camera_valid_fraction": 0.7325, "cmh_p90_m_s": 0.13554771702040083, "rhcm_p90_m_s": 0.5588012599910649, "rham_p90_deg_s": 153.91361168703025, "wms_p90": 0.9739366538798744, "gt_coverage": 0.8898163606010016, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "6f41cbd2cdc93165d1adb01c06a1e8ca3130b526863bd509f75155cc7c7551a8", "metadata_sha256": "2ae5f9b7bafd31943eadcb84bb56aa27a5697a71ca1d2bd176c6f0b3cfaa777a", "source_file": "long20_20260918/local/train/previews/da7beb32799140c1b74bb483/reference.npz", "window_s": [ 9.266666666666667, 29.233333333333334 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.654751537338637e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6b871d9b21dacc7ca1a62a0ed9a985e832f535fe112dc7a830cc0d6a35bc314d", "mano_sha256": "1d82aaf6969ae0b64ae1f057f632c66b5bcb5fd6532f3973ef12d06d495433c2", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "dc5ee74bb1063921c5f44af1", "sample_id": "bcdc7c198318b1319a3e1f67", "file": "20_seconds/ARCTIC/bcdc7c198318b1319a3e1f67_MANO_GT_20.00s.mp4", "recording_id": "s08/capsulemachine_use_04", "source_first_frame": "3", "source_last_frame": "602", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 16/600 output frames; may include genuine out-of-view frames", "id": "dc5ee74bb1063921c5f44af1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/bcdc7c198318b1319a3e1f67_MANO_GT_20.00s.mp4", "bytes": 4825175, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.020014160197568166, "p90": 0.04768961670526313, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 4.274721222893957, "p90": 10.395453397914869, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08355368441554775, "p90": 0.22286058371015163, "valid_pairs": 582, "valid_duration_s": 19.400000000000002, "coverage": 0.9716193656093489 }, "cmh_m_s": { "median": 0.03005756345318444, "p90": 0.06988720332359329, "valid_pairs": 582, "valid_duration_s": 19.400000000000002, "coverage": 0.9716193656093489 }, "rhcm_continuous_estimate_m_s": { "median": 0.08398124994886647, "p90": 0.22837691805582683, "valid_pairs": 582, "valid_duration_s": 19.400000000000002, "coverage": 0.9716193656093489 }, "rhcm_discrete_m_s": { "median": 0.08398282009625008, "p90": 0.2283249413313935, "valid_pairs": 582, "valid_duration_s": 19.400000000000002, "coverage": 0.9716193656093489 }, "wms": { "median": 0.7614828569990295, "p90": 0.94939466515127, "valid_pairs": 578, "valid_duration_s": 19.26666666666667, "coverage": 0.9649415692821369 }, "rham_deg_s": { "median": 46.36223935262895, "p90": 132.07187690253636, "valid_pairs": 582, "valid_duration_s": 19.400000000000002, "coverage": 0.9716193656093489 }, "speed_spearman": 0.31600650119151275, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.06973169436016169, "p90": 0.2086008449640571, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.973288814691152 }, "cmh_m_s": { "median": 0.029796551885655727, "p90": 0.07072792041545754, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.973288814691152 }, "rhcm_continuous_estimate_m_s": { "median": 0.07371307977207658, "p90": 0.2015658295369519, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.973288814691152 }, "rhcm_discrete_m_s": { "median": 0.07376196047209385, "p90": 0.20157094070574794, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.973288814691152 }, "wms": { "median": 0.7756880132750068, "p90": 0.9431189324704098, "valid_pairs": 582, "valid_duration_s": 19.4, "coverage": 0.9716193656093488 }, "rham_deg_s": { "median": 43.27961148553655, "p90": 161.34982876474783, "valid_pairs": 583, "valid_duration_s": 19.433333333333337, "coverage": 0.973288814691152 }, "speed_spearman": 0.22230273170804965, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "low", "relative_both_class": "low", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04768961670526313, "camera_angular_speed_p90_deg_s": 10.395453397914869, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.22286058371015163, "hand_camera_speed_p90_m_s": 0.2283249413313935, "hand_world_valid_fraction": 0.9725, "hand_camera_valid_fraction": 0.9725, "cmh_p90_m_s": 0.07072792041545754, "rhcm_p90_m_s": 0.2283249413313935, "rham_p90_deg_s": 161.34982876474783, "wms_p90": 0.94939466515127, "gt_coverage": 0.973288814691152, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "3316a8eaef21f84515e5ba8a8b7a1c53d785702a93c25a0e2cbf6f10d9ea2118", "metadata_sha256": "97ec939134625c61055643a220461716e34aec4ff7f709fe30834cbdf0a6d244", "source_file": "long20_20260918/local/train/previews/dc5ee74bb1063921c5f44af1/reference.npz", "window_s": [ 0.06666666666666667, 20.033333333333335 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.75098925995612e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "98a084c50b475bab1a2ac318826e832b490ed018fbd1d994d3814c918230f99d", "mano_sha256": "ec26c677c15e60a9e8e2bd055af6b98c0a2f85fcf3cc6b01c3c779c92eb2fd6a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "df77db1790f2b3b253ada60d", "sample_id": "febb75fcec66f1f0a9dc6815", "file": "20_seconds/ARCTIC/febb75fcec66f1f0a9dc6815_MANO_GT_20.00s.mp4", "recording_id": "s05/espressomachine_use_02", "source_first_frame": "112", "source_last_frame": "711", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 74/600 output frames; may include genuine out-of-view frames", "id": "df77db1790f2b3b253ada60d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/febb75fcec66f1f0a9dc6815_MANO_GT_20.00s.mp4", "bytes": 4826342, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.06386291297929368, "p90": 0.169159619483606, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 12.011613271847034, "p90": 43.523221626629145, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.012886815000746987, "p90": 0.47988155511937597, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8731218697829717 }, "cmh_m_s": { "median": 0.07205680566598534, "p90": 0.22187144624172297, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8731218697829717 }, "rhcm_continuous_estimate_m_s": { "median": 0.07786469870621067, "p90": 0.5425430669923615, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8731218697829717 }, "rhcm_discrete_m_s": { "median": 0.07781592580768039, "p90": 0.543071429675511, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8731218697829717 }, "wms": { "median": 0.8562460666481353, "p90": 0.9725842431821439, "valid_pairs": 468, "valid_duration_s": 15.599999999999998, "coverage": 0.7813021702838061 }, "rham_deg_s": { "median": 15.654091807449191, "p90": 95.80386872468112, "valid_pairs": 523, "valid_duration_s": 17.433333333333337, "coverage": 0.8731218697829717 }, "speed_spearman": 0.6909794449614479, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.10289516445985608, "p90": 0.34151924358853836, "valid_pairs": 470, "valid_duration_s": 15.666666666666664, "coverage": 0.7846410684474122 }, "cmh_m_s": { "median": 0.06576819993925762, "p90": 0.22276597460791955, "valid_pairs": 470, "valid_duration_s": 15.666666666666664, "coverage": 0.7846410684474122 }, "rhcm_continuous_estimate_m_s": { "median": 0.12665711452928463, "p90": 0.45081115211134853, "valid_pairs": 470, "valid_duration_s": 15.666666666666664, "coverage": 0.7846410684474122 }, "rhcm_discrete_m_s": { "median": 0.12692101182703774, "p90": 0.4484529040471814, "valid_pairs": 470, "valid_duration_s": 15.666666666666664, "coverage": 0.7846410684474122 }, "wms": { "median": 0.7335130223896629, "p90": 0.9529906882081647, "valid_pairs": 465, "valid_duration_s": 15.499999999999998, "coverage": 0.7762938230383971 }, "rham_deg_s": { "median": 27.826953118310872, "p90": 105.65243977929748, "valid_pairs": 470, "valid_duration_s": 15.666666666666664, "coverage": 0.7846410684474122 }, "speed_spearman": 0.5503009606874347, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.169159619483606, "camera_angular_speed_p90_deg_s": 43.523221626629145, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.47988155511937597, "hand_camera_speed_p90_m_s": 0.543071429675511, "hand_world_valid_fraction": 0.8325, "hand_camera_valid_fraction": 0.8325, "cmh_p90_m_s": 0.22276597460791955, "rhcm_p90_m_s": 0.543071429675511, "rham_p90_deg_s": 105.65243977929748, "wms_p90": 0.9725842431821439, "gt_coverage": 0.8731218697829717, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "c98204c0cacc54556438c550ae5488b6cf95d26ed09c3d7e0e9b899fb305c2c6", "metadata_sha256": "16a8919881ad463cb216a228536d4f3864ac6f2f899d6de92d1bfb39e7c90a2b", "source_file": "long20_20260918/local/benchmark/previews/df77db1790f2b3b253ada60d/reference.npz", "window_s": [ 3.7, 23.666666666666668 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.660953740672398e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e32bd46921d0f388f005d99a6b0fc7f3cb50558980f69a832062c597f454db79", "mano_sha256": "874f631e514e5e931746dee8b483564bb49bea0cc324926362f009793ccf1a1b", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "e17dfbc87ca19c2aa28192cc", "sample_id": "0054f543533f9f604f62dc2b", "file": "20_seconds/ARCTIC/0054f543533f9f604f62dc2b_MANO_GT_20.00s.mp4", "recording_id": "s05/waffleiron_use_02", "source_first_frame": "76", "source_last_frame": "675", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "", "id": "e17dfbc87ca19c2aa28192cc", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/0054f543533f9f604f62dc2b_MANO_GT_20.00s.mp4", "bytes": 6288649, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08326063412304208, "p90": 0.2228983055128565, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 19.42583559547096, "p90": 50.9667548724227, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.13034024359151614, "p90": 0.4545004178531045, "valid_pairs": 584, "valid_duration_s": 19.46666666666667, "coverage": 0.9749582637729551 }, "cmh_m_s": { "median": 0.10001858108665942, "p90": 0.22666406393475275, "valid_pairs": 584, "valid_duration_s": 19.46666666666667, "coverage": 0.9749582637729551 }, "rhcm_continuous_estimate_m_s": { "median": 0.16960658803263878, "p90": 0.47556583131924923, "valid_pairs": 584, "valid_duration_s": 19.46666666666667, "coverage": 0.9749582637729551 }, "rhcm_discrete_m_s": { "median": 0.16929960121203977, "p90": 0.476426984502006, "valid_pairs": 584, "valid_duration_s": 19.46666666666667, "coverage": 0.9749582637729551 }, "wms": { "median": 0.6975028351029753, "p90": 0.9407374315515716, "valid_pairs": 584, "valid_duration_s": 19.46666666666667, "coverage": 0.9749582637729551 }, "rham_deg_s": { "median": 81.8313689791118, "p90": 208.060903223975, "valid_pairs": 584, "valid_duration_s": 19.46666666666667, "coverage": 0.9749582637729551 }, "speed_spearman": 0.273919825328458, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.12554577730309735, "p90": 0.3733491513151437, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "cmh_m_s": { "median": 0.09306453386211884, "p90": 0.2220103903263598, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "rhcm_continuous_estimate_m_s": { "median": 0.15572275172704259, "p90": 0.39197516723691994, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "rhcm_discrete_m_s": { "median": 0.15574174766601287, "p90": 0.39382375048881946, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "wms": { "median": 0.7316042811870751, "p90": 0.9527387833892733, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "rham_deg_s": { "median": 63.73226619858537, "p90": 176.44721851480008, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "speed_spearman": 0.20379805412035082, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.2228983055128565, "camera_angular_speed_p90_deg_s": 50.9667548724227, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4545004178531045, "hand_camera_speed_p90_m_s": 0.476426984502006, "hand_world_valid_fraction": 0.9675, "hand_camera_valid_fraction": 0.9675, "cmh_p90_m_s": 0.22666406393475275, "rhcm_p90_m_s": 0.476426984502006, "rham_p90_deg_s": 208.060903223975, "wms_p90": 0.9527387833892733, "gt_coverage": 0.9749582637729551, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "608c05577c091a14fecd628a9f4e7a5001ae4d83286acdef779b973c91f251bb", "metadata_sha256": "4d4265d627bb5f26cff94fa91c662dc92e784e155207beaf36985e106a0846bc", "source_file": "long20_20260918/local/benchmark/previews/e17dfbc87ca19c2aa28192cc/reference.npz", "window_s": [ 2.5, 22.466666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.611102534756203e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "53a5c8b42111e04f4e34ef7b49651e5e5ee3f36d7e290aa4e2b0820d281f0079", "mano_sha256": "26d287e95c9ee77b2c468b55aa548293733fd2d2a2c3a78d3b0324d588e16b05", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e2b4a93e1b7abfbf7ea23caa", "sample_id": "0d41fe798819b9576794c7f9", "file": "20_seconds/ARCTIC/0d41fe798819b9576794c7f9_MANO_GT_20.00s.mp4", "recording_id": "s08/laptop_use_03", "source_first_frame": "229", "source_last_frame": "828", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 42/600 output frames; may include genuine out-of-view frames", "id": "e2b4a93e1b7abfbf7ea23caa", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/0d41fe798819b9576794c7f9_MANO_GT_20.00s.mp4", "bytes": 5244012, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.024769110027210627, "p90": 0.0638797281546577, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 5.629838222729804, "p90": 15.536041271251534, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.11145315644047038, "p90": 0.3101337555797663, "valid_pairs": 553, "valid_duration_s": 18.433333333333334, "coverage": 0.9232053422370617 }, "cmh_m_s": { "median": 0.03804093429868361, "p90": 0.09339689682069613, "valid_pairs": 553, "valid_duration_s": 18.433333333333334, "coverage": 0.9232053422370617 }, "rhcm_continuous_estimate_m_s": { "median": 0.09664352271382153, "p90": 0.3251292608332099, "valid_pairs": 553, "valid_duration_s": 18.433333333333334, "coverage": 0.9232053422370617 }, "rhcm_discrete_m_s": { "median": 0.0966270932051908, "p90": 0.32519487090143967, "valid_pairs": 553, "valid_duration_s": 18.433333333333334, "coverage": 0.9232053422370617 }, "wms": { "median": 0.7060460635422267, "p90": 0.9525220998637459, "valid_pairs": 551, "valid_duration_s": 18.366666666666667, "coverage": 0.9198664440734557 }, "rham_deg_s": { "median": 40.70490897862542, "p90": 130.02481053934488, "valid_pairs": 553, "valid_duration_s": 18.433333333333334, "coverage": 0.9232053422370617 }, "speed_spearman": 0.25528090809309706, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.07735973599073498, "p90": 0.28093910453612025, "valid_pairs": 557, "valid_duration_s": 18.56666666666667, "coverage": 0.9298831385642738 }, "cmh_m_s": { "median": 0.03690189980412971, "p90": 0.08946977371868181, "valid_pairs": 557, "valid_duration_s": 18.56666666666667, "coverage": 0.9298831385642738 }, "rhcm_continuous_estimate_m_s": { "median": 0.06534597203134968, "p90": 0.2887892239893375, "valid_pairs": 557, "valid_duration_s": 18.56666666666667, "coverage": 0.9298831385642738 }, "rhcm_discrete_m_s": { "median": 0.06531675986721408, "p90": 0.2887890712366177, "valid_pairs": 557, "valid_duration_s": 18.56666666666667, "coverage": 0.9298831385642738 }, "wms": { "median": 0.6655254625037459, "p90": 0.9422250207600099, "valid_pairs": 556, "valid_duration_s": 18.53333333333334, "coverage": 0.928213689482471 }, "rham_deg_s": { "median": 29.649574190757335, "p90": 137.4085272602132, "valid_pairs": 557, "valid_duration_s": 18.56666666666667, "coverage": 0.9298831385642738 }, "speed_spearman": 0.2959837029895398, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0638797281546577, "camera_angular_speed_p90_deg_s": 15.536041271251534, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3101337555797663, "hand_camera_speed_p90_m_s": 0.32519487090143967, "hand_world_valid_fraction": 0.9275, "hand_camera_valid_fraction": 0.9275, "cmh_p90_m_s": 0.09339689682069613, "rhcm_p90_m_s": 0.32519487090143967, "rham_p90_deg_s": 137.4085272602132, "wms_p90": 0.9525220998637459, "gt_coverage": 0.9298831385642738, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "632b1370277807aa3b369d4a791dc31789ff0bb1261bf100ac28ec88ec8e11d5", "metadata_sha256": "2b086aaef379626965acd138b02ef717375f7583346a210f870d731db45a0374", "source_file": "long20_20260918/local/train/previews/e2b4a93e1b7abfbf7ea23caa/reference.npz", "window_s": [ 7.6, 27.566666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.611102534756203e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3926a5e30c8f1c7f125a0a71c57863643287c2d6bc4f396748cadd9500b4462d", "mano_sha256": "9a44021e5704fa19650dee2110ad234bc9459e8ba69ed723130bf4268cb04043", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e2fa62b38b4f9213dc4feab3", "sample_id": "5b26468e07094f955a292777", "file": "20_seconds/ARCTIC/5b26468e07094f955a292777_MANO_GT_20.00s.mp4", "recording_id": "s09/microwave_use_04", "source_first_frame": "2", "source_last_frame": "601", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 32/600 output frames; may include genuine out-of-view frames", "id": "e2fa62b38b4f9213dc4feab3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/5b26468e07094f955a292777_MANO_GT_20.00s.mp4", "bytes": 3937134, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.021343290580815976, "p90": 0.04792660138565022, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 4.812696654866846, "p90": 8.606449696480103, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08959417547430519, "p90": 0.41537131411236744, "valid_pairs": 562, "valid_duration_s": 18.733333333333334, "coverage": 0.938230383973289 }, "cmh_m_s": { "median": 0.04307154037233798, "p90": 0.09517916618243703, "valid_pairs": 562, "valid_duration_s": 18.733333333333334, "coverage": 0.938230383973289 }, "rhcm_continuous_estimate_m_s": { "median": 0.10416475533297037, "p90": 0.4015186324988817, "valid_pairs": 562, "valid_duration_s": 18.733333333333334, "coverage": 0.938230383973289 }, "rhcm_discrete_m_s": { "median": 0.10419392150525303, "p90": 0.4014738394037926, "valid_pairs": 562, "valid_duration_s": 18.733333333333334, "coverage": 0.938230383973289 }, "wms": { "median": 0.7574269157100136, "p90": 0.9417575331123705, "valid_pairs": 561, "valid_duration_s": 18.7, "coverage": 0.9365609348914858 }, "rham_deg_s": { "median": 37.68502888486555, "p90": 122.82161854487344, "valid_pairs": 562, "valid_duration_s": 18.733333333333334, "coverage": 0.938230383973289 }, "speed_spearman": 0.40664341735712806, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.007572959668248837, "p90": 0.024058310891005266, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "cmh_m_s": { "median": 0.044434933582967746, "p90": 0.09165353155964963, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "rhcm_continuous_estimate_m_s": { "median": 0.045755341317217965, "p90": 0.09621061300039983, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "rhcm_discrete_m_s": { "median": 0.04564339052612075, "p90": 0.09628758867284727, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "wms": { "median": 0.8610560337474715, "p90": 0.9569234589688952, "valid_pairs": 545, "valid_duration_s": 18.16666666666667, "coverage": 0.9098497495826381 }, "rham_deg_s": { "median": 7.466049551716189, "p90": 27.878527896876694, "valid_pairs": 567, "valid_duration_s": 18.9, "coverage": 0.9465776293823038 }, "speed_spearman": 0.22359856034518852, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.04792660138565022, "camera_angular_speed_p90_deg_s": 8.606449696480103, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.41537131411236744, "hand_camera_speed_p90_m_s": 0.4014738394037926, "hand_world_valid_fraction": 0.9425, "hand_camera_valid_fraction": 0.9425, "cmh_p90_m_s": 0.09517916618243703, "rhcm_p90_m_s": 0.4014738394037926, "rham_p90_deg_s": 122.82161854487344, "wms_p90": 0.9569234589688952, "gt_coverage": 0.9465776293823038, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "f2d3d5f938143623ec3c82cedad07fb16238412bf6a1f65622de8fb6e52c9113", "metadata_sha256": "11a81b85ddf3c94d56631a58815d02e4445510be5d062e8a19c2382fcf12e607", "source_file": "long20_20260918/local/train/previews/e2fa62b38b4f9213dc4feab3/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.743382398173679e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "26e7bd07f5f8c6feff6103c8ef9a91c545364cc3c0605de3d162f56f9ec65ee9", "mano_sha256": "55d6f9675a21f4c56fb16c683d6b119aceff0306c8cb01fa020444f97c3b039c", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "e4b32b9633de106b13367459", "sample_id": "ef6cb1ec2c41ee45f4183086", "file": "20_seconds/ARCTIC/ef6cb1ec2c41ee45f4183086_MANO_GT_20.00s.mp4", "recording_id": "s02/laptop_use_02", "source_first_frame": "142", "source_last_frame": "741", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 57/600 output frames; may include genuine out-of-view frames", "id": "e4b32b9633de106b13367459", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/ef6cb1ec2c41ee45f4183086_MANO_GT_20.00s.mp4", "bytes": 4744276, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.024279184964693366, "p90": 0.054453705344644755, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "camera_rotation": { "median": 4.762508715002649, "p90": 11.151419269203474, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0000000000000002 }, "sides": { "left": { "whm_m_s": { "median": 0.024748945632096807, "p90": 0.21627300271908484, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.898163606010017 }, "cmh_m_s": { "median": 0.028302238286978698, "p90": 0.051780114670502524, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.898163606010017 }, "rhcm_continuous_estimate_m_s": { "median": 0.04516760298065174, "p90": 0.20368259699913363, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.898163606010017 }, "rhcm_discrete_m_s": { "median": 0.04517968413276008, "p90": 0.20372107622371463, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.898163606010017 }, "wms": { "median": 0.7850239680198379, "p90": 0.960295242613382, "valid_pairs": 512, "valid_duration_s": 17.066666666666666, "coverage": 0.8547579298831386 }, "rham_deg_s": { "median": 23.796591611652325, "p90": 117.53290274039347, "valid_pairs": 538, "valid_duration_s": 17.933333333333337, "coverage": 0.898163606010017 }, "speed_spearman": 0.24060923218929134, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.14696971467442754, "p90": 0.47149754489291285, "valid_pairs": 542, "valid_duration_s": 18.06666666666667, "coverage": 0.904841402337229 }, "cmh_m_s": { "median": 0.03413079009216863, "p90": 0.08127618855746953, "valid_pairs": 542, "valid_duration_s": 18.06666666666667, "coverage": 0.904841402337229 }, "rhcm_continuous_estimate_m_s": { "median": 0.14229141997493294, "p90": 0.4498061352817182, "valid_pairs": 542, "valid_duration_s": 18.06666666666667, "coverage": 0.904841402337229 }, "rhcm_discrete_m_s": { "median": 0.14237778365074016, "p90": 0.44980556576817265, "valid_pairs": 542, "valid_duration_s": 18.06666666666667, "coverage": 0.904841402337229 }, "wms": { "median": 0.7915392778484394, "p90": 0.9593995418957676, "valid_pairs": 542, "valid_duration_s": 18.06666666666667, "coverage": 0.904841402337229 }, "rham_deg_s": { "median": 47.92030404697579, "p90": 165.67419556957478, "valid_pairs": 542, "valid_duration_s": 18.06666666666667, "coverage": 0.904841402337229 }, "speed_spearman": 0.2887885260017008, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.054453705344644755, "camera_angular_speed_p90_deg_s": 11.151419269203474, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.47149754489291285, "hand_camera_speed_p90_m_s": 0.44980556576817265, "hand_world_valid_fraction": 0.9016666666666666, "hand_camera_valid_fraction": 0.9016666666666666, "cmh_p90_m_s": 0.08127618855746953, "rhcm_p90_m_s": 0.44980556576817265, "rham_p90_deg_s": 165.67419556957478, "wms_p90": 0.960295242613382, "gt_coverage": 0.904841402337229, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "d365b4a7f21bf6e6e95a26060ebbd3e11b300050202b220c4636284bfef40df7", "metadata_sha256": "5975716104ba71bad8b1f52fcb2e6e6e8fe1d23d92cc3196df34a2400a5b7770", "source_file": "long20_20260918/local/train/previews/e4b32b9633de106b13367459/reference.npz", "window_s": [ 4.666666666666667, 24.633333333333333 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.44522679167363e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5e6d3d79beaee76e02c61214f48a6d30d3827873bf1980e00f0fd143e67d9c01", "mano_sha256": "cbafd75eeadbbd16fd942c0decd1f3ac87871aeecd9fea876e9b93240818a274", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "ecfef0d44e3c152aba84283d", "sample_id": "94fed2f456d91fc688485f2d", "file": "20_seconds/ARCTIC/94fed2f456d91fc688485f2d_MANO_GT_20.00s.mp4", "recording_id": "s05/notebook_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 29/600 output frames; may include genuine out-of-view frames", "id": "ecfef0d44e3c152aba84283d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/94fed2f456d91fc688485f2d_MANO_GT_20.00s.mp4", "bytes": 5894097, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.07045146231674611, "p90": 0.15802166934770295, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 18.056106626497403, "p90": 37.73962330286099, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.11927742896104512, "p90": 0.3034603283635057, "valid_pairs": 298, "valid_duration_s": 9.933333333333335, "coverage": 0.4974958263772956 }, "cmh_m_s": { "median": 0.08506283607513729, "p90": 0.20881958349167815, "valid_pairs": 298, "valid_duration_s": 9.933333333333335, "coverage": 0.4974958263772956 }, "rhcm_continuous_estimate_m_s": { "median": 0.15102192407254877, "p90": 0.34056181276826986, "valid_pairs": 298, "valid_duration_s": 9.933333333333335, "coverage": 0.4974958263772956 }, "rhcm_discrete_m_s": { "median": 0.15086847175252094, "p90": 0.34039765963193447, "valid_pairs": 298, "valid_duration_s": 9.933333333333335, "coverage": 0.4974958263772956 }, "wms": { "median": 0.7281695430621131, "p90": 0.9554826831558058, "valid_pairs": 298, "valid_duration_s": 9.933333333333335, "coverage": 0.4974958263772956 }, "rham_deg_s": { "median": 83.07017098638516, "p90": 229.69064981354123, "valid_pairs": 298, "valid_duration_s": 9.933333333333335, "coverage": 0.4974958263772956 }, "speed_spearman": 0.20416936757759635, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.08259177549945687, "p90": 0.20437621276391557, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "cmh_m_s": { "median": 0.08420955077919094, "p90": 0.20084616795048607, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "rhcm_continuous_estimate_m_s": { "median": 0.14178135897347074, "p90": 0.3100561588335257, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "rhcm_discrete_m_s": { "median": 0.14154670081030563, "p90": 0.3108828709702441, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "wms": { "median": 0.7782523861100631, "p90": 0.9405861765771241, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "rham_deg_s": { "median": 57.0963823290311, "p90": 146.54344064582534, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285476 }, "speed_spearman": 0.12516329849339297, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.15802166934770295, "camera_angular_speed_p90_deg_s": 37.73962330286099, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3034603283635057, "hand_camera_speed_p90_m_s": 0.34039765963193447, "hand_world_valid_fraction": 0.6891666666666667, "hand_camera_valid_fraction": 0.6891666666666667, "cmh_p90_m_s": 0.20881958349167815, "rhcm_p90_m_s": 0.34039765963193447, "rham_p90_deg_s": 229.69064981354123, "wms_p90": 0.9554826831558058, "gt_coverage": 0.8597662771285476, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "93571fc251537abb5e69c0b6d72e2c8b03ca87fe7b677d544bf0f4b25b8d663e", "metadata_sha256": "1d595d0fd759fd1e3409684df1f7a8a603ef528a25c3f2d5cb34b1c997b7302b", "source_file": "long20_20260918/local/benchmark/previews/ecfef0d44e3c152aba84283d/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.579669976578771e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "438e9be72df4e90331e704edbea5e314e8e13b73708e181839c1010843d57d58", "mano_sha256": "d99edc3b026b3340d07112ccc184846811d75334570b61425ce9fc3406c51ff6", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "f2643ea09abe30143f28f403", "sample_id": "bed91c7539053ef8aeaf4255", "file": "20_seconds/ARCTIC/bed91c7539053ef8aeaf4255_MANO_GT_20.00s.mp4", "recording_id": "s02/phone_use_02", "source_first_frame": "4", "source_last_frame": "603", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 55/600 output frames; may include genuine out-of-view frames", "id": "f2643ea09abe30143f28f403", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/bed91c7539053ef8aeaf4255_MANO_GT_20.00s.mp4", "bytes": 4887416, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.022106220526527033, "p90": 0.09506994114232475, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.7744529020217548, "p90": 15.087808587990532, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.11179271493827045, "p90": 0.36919809216306426, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285474 }, "cmh_m_s": { "median": 0.020347882426946347, "p90": 0.048316605054445544, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285474 }, "rhcm_continuous_estimate_m_s": { "median": 0.10959245378697131, "p90": 0.3565581636538092, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285474 }, "rhcm_discrete_m_s": { "median": 0.10965748947400669, "p90": 0.3563621208004699, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285474 }, "wms": { "median": 0.8308204428044875, "p90": 0.965408732738859, "valid_pairs": 507, "valid_duration_s": 16.9, "coverage": 0.8464106844741234 }, "rham_deg_s": { "median": 50.125255508758116, "p90": 126.19256530590896, "valid_pairs": 515, "valid_duration_s": 17.166666666666664, "coverage": 0.8597662771285474 }, "speed_spearman": 0.45501096856322804, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.032980655010769036, "p90": 0.19567817409361257, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008346 }, "cmh_m_s": { "median": 0.020626893839227344, "p90": 0.06172469761073938, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008346 }, "rhcm_continuous_estimate_m_s": { "median": 0.03832959978984514, "p90": 0.14720002244906485, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008346 }, "rhcm_discrete_m_s": { "median": 0.03833241509831361, "p90": 0.14721422211872462, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008346 }, "wms": { "median": 0.7032886110674966, "p90": 0.9236757769090977, "valid_pairs": 523, "valid_duration_s": 17.43333333333333, "coverage": 0.8731218697829713 }, "rham_deg_s": { "median": 23.64897209898085, "p90": 89.77362143122497, "valid_pairs": 544, "valid_duration_s": 18.133333333333333, "coverage": 0.9081803005008346 }, "speed_spearman": 0.4798865808171309, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "low", "relative_class": "low", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.09506994114232475, "camera_angular_speed_p90_deg_s": 15.087808587990532, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.36919809216306426, "hand_camera_speed_p90_m_s": 0.3563621208004699, "hand_world_valid_fraction": 0.885, "hand_camera_valid_fraction": 0.885, "cmh_p90_m_s": 0.06172469761073938, "rhcm_p90_m_s": 0.3563621208004699, "rham_p90_deg_s": 126.19256530590896, "wms_p90": 0.965408732738859, "gt_coverage": 0.9081803005008346, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "5af6318f887dd961009437d0cbf1868d66b90db15d00b130609b158da6932ec0", "metadata_sha256": "f034bfa185222ee5d81f3ca878a238807b98186eb3af0ba3ca178274e512efcb", "source_file": "long20_20260918/local/train/previews/f2643ea09abe30143f28f403/reference.npz", "window_s": [ 0.06666666666666667, 20.033333333333335 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.579669976578771e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f80774206866fdb7fe743517372c39870d487ae7c8f1b7425690f1237d2242aa", "mano_sha256": "69ce98e9d3aacb1b8008d3a813b323637ecb9d707f3736405760e61d4a0eeabe", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "f2d4ad897e674d453288ae0a", "sample_id": "b8db97ba3b42a83fe141c7e4", "file": "20_seconds/ARCTIC/b8db97ba3b42a83fe141c7e4_MANO_GT_20.00s.mp4", "recording_id": "s05/waffleiron_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 22/600 output frames; may include genuine out-of-view frames", "id": "f2d4ad897e674d453288ae0a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/b8db97ba3b42a83fe141c7e4_MANO_GT_20.00s.mp4", "bytes": 5770824, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.09236707843829198, "p90": 0.19326735359455605, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 16.581274301136602, "p90": 34.227717849975384, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.12267946605648536, "p90": 0.3613274690799948, "valid_pairs": 558, "valid_duration_s": 18.599999999999998, "coverage": 0.9315525876460767 }, "cmh_m_s": { "median": 0.1008493260830104, "p90": 0.20143802879571895, "valid_pairs": 558, "valid_duration_s": 18.599999999999998, "coverage": 0.9315525876460767 }, "rhcm_continuous_estimate_m_s": { "median": 0.16124163375371825, "p90": 0.3975659137899345, "valid_pairs": 558, "valid_duration_s": 18.599999999999998, "coverage": 0.9315525876460767 }, "rhcm_discrete_m_s": { "median": 0.16107928093519552, "p90": 0.3975461956472395, "valid_pairs": 558, "valid_duration_s": 18.599999999999998, "coverage": 0.9315525876460767 }, "wms": { "median": 0.7495282570927619, "p90": 0.9543343465660933, "valid_pairs": 558, "valid_duration_s": 18.599999999999998, "coverage": 0.9315525876460767 }, "rham_deg_s": { "median": 64.16563518023315, "p90": 159.72436129572597, "valid_pairs": 558, "valid_duration_s": 18.599999999999998, "coverage": 0.9315525876460767 }, "speed_spearman": 0.22801353563083496, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.11661607820260386, "p90": 0.40772996695946, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "cmh_m_s": { "median": 0.09966267070378156, "p90": 0.1983251114021525, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rhcm_continuous_estimate_m_s": { "median": 0.15743998802401646, "p90": 0.40513747096605346, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rhcm_discrete_m_s": { "median": 0.15663231217874013, "p90": 0.4050224989660758, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "wms": { "median": 0.7325651590136936, "p90": 0.9427763535272262, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "rham_deg_s": { "median": 57.840967024700056, "p90": 189.5849618769133, "valid_pairs": 577, "valid_duration_s": 19.233333333333334, "coverage": 0.963272120200334 }, "speed_spearman": 0.23658132587319775, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.19326735359455605, "camera_angular_speed_p90_deg_s": 34.227717849975384, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.40772996695946, "hand_camera_speed_p90_m_s": 0.4050224989660758, "hand_world_valid_fraction": 0.9483333333333334, "hand_camera_valid_fraction": 0.9483333333333334, "cmh_p90_m_s": 0.20143802879571895, "rhcm_p90_m_s": 0.4050224989660758, "rham_p90_deg_s": 189.5849618769133, "wms_p90": 0.9543343465660933, "gt_coverage": 0.963272120200334, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "13ab56003a69aef7e0395a09c3bd04a9990f2fd61b762ef91199d8667680f2bd", "metadata_sha256": "a901aaa998c37663e36432e217e1deba15d2bf02b222d1a2ff9b626eaf8ada9c", "source_file": "long20_20260918/local/benchmark/previews/f2d4ad897e674d453288ae0a/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.742874840267547e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "1bddc2295a2dbb3974cae3c5808e2fc76970d0d44aa503eb6d7d67140dab9a2b", "mano_sha256": "92e67084c498bf519e8dd2b698c693ef1ec4f345cc6d85ecdbc9a08ecd534a51", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "f6c7738dbf77c081d33b5b4b", "sample_id": "fb53e06f5ef2b8e05c4f77ca", "file": "20_seconds/ARCTIC/fb53e06f5ef2b8e05c4f77ca_MANO_GT_20.00s.mp4", "recording_id": "s05/laptop_grab_01", "source_first_frame": "229", "source_last_frame": "828", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 68/600 output frames; may include genuine out-of-view frames", "id": "f6c7738dbf77c081d33b5b4b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/fb53e06f5ef2b8e05c4f77ca_MANO_GT_20.00s.mp4", "bytes": 5839466, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.08633434839675735, "p90": 0.1857709588351933, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 21.798905711467974, "p90": 49.906213918581805, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.2743050273403521, "p90": 0.5744250865908145, "valid_pairs": 284, "valid_duration_s": 9.466666666666667, "coverage": 0.47412353923205336 }, "cmh_m_s": { "median": 0.09565636318344552, "p90": 0.22255084421047136, "valid_pairs": 284, "valid_duration_s": 9.466666666666667, "coverage": 0.47412353923205336 }, "rhcm_continuous_estimate_m_s": { "median": 0.30612829542548525, "p90": 0.5955904076322042, "valid_pairs": 284, "valid_duration_s": 9.466666666666667, "coverage": 0.47412353923205336 }, "rhcm_discrete_m_s": { "median": 0.3061680028219388, "p90": 0.5960542934191881, "valid_pairs": 284, "valid_duration_s": 9.466666666666667, "coverage": 0.47412353923205336 }, "wms": { "median": 0.7632270728507953, "p90": 0.9230220269415961, "valid_pairs": 284, "valid_duration_s": 9.466666666666667, "coverage": 0.47412353923205336 }, "rham_deg_s": { "median": 107.00235983188693, "p90": 242.78895298234116, "valid_pairs": 284, "valid_duration_s": 9.466666666666667, "coverage": 0.47412353923205336 }, "speed_spearman": -0.0301126079519707, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.21581738992365423, "p90": 0.5403759074000105, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "cmh_m_s": { "median": 0.12478011461560748, "p90": 0.24186106863151177, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "rhcm_continuous_estimate_m_s": { "median": 0.21125586172345037, "p90": 0.5021665148226464, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "rhcm_discrete_m_s": { "median": 0.21119784205614878, "p90": 0.5027738575045139, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "wms": { "median": 0.6428937663917249, "p90": 0.8794115776954469, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "rham_deg_s": { "median": 89.58502500159531, "p90": 226.19249912774276, "valid_pairs": 511, "valid_duration_s": 17.03333333333333, "coverage": 0.8530884808013354 }, "speed_spearman": 0.20401129201680665, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.1857709588351933, "camera_angular_speed_p90_deg_s": 49.906213918581805, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5744250865908145, "hand_camera_speed_p90_m_s": 0.5960542934191881, "hand_world_valid_fraction": 0.6683333333333333, "hand_camera_valid_fraction": 0.6683333333333333, "cmh_p90_m_s": 0.24186106863151177, "rhcm_p90_m_s": 0.5960542934191881, "rham_p90_deg_s": 242.78895298234116, "wms_p90": 0.9230220269415961, "gt_coverage": 0.8530884808013354, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "2bd065e2ee61f69aa1392becc7304872f26a3611279775780084da88a0bf692f", "metadata_sha256": "00288800db73a8ec981647b7df614c22d49e47588d14f021c9ce730438760224", "source_file": "long20_20260918/local/benchmark/previews/f6c7738dbf77c081d33b5b4b/reference.npz", "window_s": [ 7.6, 27.566666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.654751537338637e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "55e80f4cacef9b4164b0eca63a385cbce3e9883f4983ff07493b0aea6b8888e5", "mano_sha256": "bf3ccbf6dbba7c786bad17b05720da70492549e6d52251254ce6d4383b014e4a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "f8d5d750c2667466f1390d78", "sample_id": "6f7cf92e1895310dd7a5a5d9", "file": "20_seconds/ARCTIC/6f7cf92e1895310dd7a5a5d9_MANO_GT_20.00s.mp4", "recording_id": "s05/espressomachine_grab_01", "source_first_frame": "289", "source_last_frame": "888", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 89/600 output frames; may include genuine out-of-view frames", "id": "f8d5d750c2667466f1390d78", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/6f7cf92e1895310dd7a5a5d9_MANO_GT_20.00s.mp4", "bytes": 5480192, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.07698812498855903, "p90": 0.21205076755050353, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 11.790762811314162, "p90": 39.63353850160563, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.19815840175758417, "p90": 0.4159992182348263, "valid_pairs": 370, "valid_duration_s": 12.333333333333332, "coverage": 0.6176961602671117 }, "cmh_m_s": { "median": 0.07873492618502341, "p90": 0.19357624824620087, "valid_pairs": 370, "valid_duration_s": 12.333333333333332, "coverage": 0.6176961602671117 }, "rhcm_continuous_estimate_m_s": { "median": 0.22507355372474597, "p90": 0.445977006033363, "valid_pairs": 370, "valid_duration_s": 12.333333333333332, "coverage": 0.6176961602671117 }, "rhcm_discrete_m_s": { "median": 0.22473272465313132, "p90": 0.4460475014231848, "valid_pairs": 370, "valid_duration_s": 12.333333333333332, "coverage": 0.6176961602671117 }, "wms": { "median": 0.7687782365643698, "p90": 0.9586164774411633, "valid_pairs": 370, "valid_duration_s": 12.333333333333332, "coverage": 0.6176961602671117 }, "rham_deg_s": { "median": 56.829349150778, "p90": 141.77200801927484, "valid_pairs": 370, "valid_duration_s": 12.333333333333332, "coverage": 0.6176961602671117 }, "speed_spearman": 0.1323455465194996, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.21368178818713776, "p90": 0.5531403405662848, "valid_pairs": 459, "valid_duration_s": 15.299999999999999, "coverage": 0.7662771285475791 }, "cmh_m_s": { "median": 0.08832137489805068, "p90": 0.24811807361446103, "valid_pairs": 459, "valid_duration_s": 15.299999999999999, "coverage": 0.7662771285475791 }, "rhcm_continuous_estimate_m_s": { "median": 0.1996499405392438, "p90": 0.4749770006813058, "valid_pairs": 459, "valid_duration_s": 15.299999999999999, "coverage": 0.7662771285475791 }, "rhcm_discrete_m_s": { "median": 0.19958771362746847, "p90": 0.4749356608980003, "valid_pairs": 459, "valid_duration_s": 15.299999999999999, "coverage": 0.7662771285475791 }, "wms": { "median": 0.6829960932236416, "p90": 0.9089067147064573, "valid_pairs": 459, "valid_duration_s": 15.299999999999999, "coverage": 0.7662771285475791 }, "rham_deg_s": { "median": 55.799594829248434, "p90": 157.33436457292237, "valid_pairs": 459, "valid_duration_s": 15.299999999999999, "coverage": 0.7662771285475791 }, "speed_spearman": 0.40343959367178295, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "high", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.21205076755050353, "camera_angular_speed_p90_deg_s": 39.63353850160563, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.5531403405662848, "hand_camera_speed_p90_m_s": 0.4749356608980003, "hand_world_valid_fraction": 0.6991666666666667, "hand_camera_valid_fraction": 0.6991666666666667, "cmh_p90_m_s": 0.24811807361446103, "rhcm_p90_m_s": 0.4749356608980003, "rham_p90_deg_s": 157.33436457292237, "wms_p90": 0.9586164774411633, "gt_coverage": 0.7662771285475791, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "65a442b1a5af3659c73db3edb6e1ba09943a61b37c8b7a611c0c98c8c47eef1a", "metadata_sha256": "a2aa48149a0bec462227911c0699818f1832fde7f1bdecd82ddf99ae26ac7682", "source_file": "long20_20260918/local/benchmark/previews/f8d5d750c2667466f1390d78/reference.npz", "window_s": [ 9.6, 29.566666666666666 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.552252871008668e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "9606d3255abac71b0172169b461c3fb5f5c1adcc71259247067c48546d8bd609", "mano_sha256": "b0b14d78ceef058315cbae90fcf690971cc9e41ea738894564ffa7b401d4e85e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "fb81de49340ce2e6c19bed4b", "sample_id": "69fd97b34b861ad7cac38f81", "file": "20_seconds/ARCTIC/69fd97b34b861ad7cac38f81_MANO_GT_20.00s.mp4", "recording_id": "s05/capsulemachine_use_01", "source_first_frame": "1", "source_last_frame": "600", "source_split": "official_p2_val", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 24/600 output frames; may include genuine out-of-view frames", "id": "fb81de49340ce2e6c19bed4b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/69fd97b34b861ad7cac38f81_MANO_GT_20.00s.mp4", "bytes": 5625299, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0739187025983285, "p90": 0.20172244629896408, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 19.300967677428698, "p90": 45.829619413478355, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08004578616751018, "p90": 0.332874521418922, "valid_pairs": 575, "valid_duration_s": 19.166666666666664, "coverage": 0.9599332220367278 }, "cmh_m_s": { "median": 0.10028967982665857, "p90": 0.2356423491013801, "valid_pairs": 575, "valid_duration_s": 19.166666666666664, "coverage": 0.9599332220367278 }, "rhcm_continuous_estimate_m_s": { "median": 0.15601397750848475, "p90": 0.378264937294734, "valid_pairs": 575, "valid_duration_s": 19.166666666666664, "coverage": 0.9599332220367278 }, "rhcm_discrete_m_s": { "median": 0.15639003666095716, "p90": 0.3776450301220099, "valid_pairs": 575, "valid_duration_s": 19.166666666666664, "coverage": 0.9599332220367278 }, "wms": { "median": 0.8228460309288752, "p90": 0.9739842030006197, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "rham_deg_s": { "median": 46.33432295458629, "p90": 137.3540321206769, "valid_pairs": 575, "valid_duration_s": 19.166666666666664, "coverage": 0.9599332220367278 }, "speed_spearman": 0.11062616775236075, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.12744723805020414, "p90": 0.33671445803172434, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "cmh_m_s": { "median": 0.09834144805400166, "p90": 0.20590826106047994, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "rhcm_continuous_estimate_m_s": { "median": 0.16505664706791684, "p90": 0.40828208880599687, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "rhcm_discrete_m_s": { "median": 0.16489874641972857, "p90": 0.40804449704051743, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "wms": { "median": 0.7862259095490336, "p90": 0.9667052221445591, "valid_pairs": 571, "valid_duration_s": 19.03333333333333, "coverage": 0.9532554257095158 }, "rham_deg_s": { "median": 54.56842482245166, "p90": 146.24546212042677, "valid_pairs": 572, "valid_duration_s": 19.066666666666663, "coverage": 0.9549248747913187 }, "speed_spearman": 0.07925531891019096, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" } }, "hand_both_class": "high", "relative_both_class": "high", "relative_pose_class": "moving", "camera_class": "high", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.20172244629896408, "camera_angular_speed_p90_deg_s": 45.829619413478355, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.33671445803172434, "hand_camera_speed_p90_m_s": 0.40804449704051743, "hand_world_valid_fraction": 0.9575, "hand_camera_valid_fraction": 0.9575, "cmh_p90_m_s": 0.2356423491013801, "rhcm_p90_m_s": 0.40804449704051743, "rham_p90_deg_s": 146.24546212042677, "wms_p90": 0.9739842030006197, "gt_coverage": 0.9599332220367278, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "f19c581b6976c99261cafb360bccfbe7e6e07ea82579813f6d32260ed6f2380a", "metadata_sha256": "54bf5e7dfab189fb1881c94fcf0a65c583745776318e4e045091cf7be3db5216", "source_file": "long20_20260918/local/benchmark/previews/fb81de49340ce2e6c19bed4b/reference.npz", "window_s": [ 0.0, 19.966666666666665 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.585973723508303e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4c3e018b1d398c7347246237be486fe99857743a5ee9055240c131d7d7c39070", "mano_sha256": "a1f2040c2b55581200c8006d2ebd9077dc297b4a1e957d94e1b91a8cfadd1b6c", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "fd57bc093781ef7f6d597bd1", "sample_id": "c432c1d263086af4623ce910", "file": "20_seconds/ARCTIC/c432c1d263086af4623ce910_MANO_GT_20.00s.mp4", "recording_id": "s04/microwave_use_02", "source_first_frame": "4", "source_last_frame": "603", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 25/600 output frames; may include genuine out-of-view frames", "id": "fd57bc093781ef7f6d597bd1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/c432c1d263086af4623ce910_MANO_GT_20.00s.mp4", "bytes": 4680334, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.029772578712687448, "p90": 0.06116824786105195, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 6.803246558062993, "p90": 13.966974060209177, "valid_pairs": 599, "valid_duration_s": 19.96666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.18572247863074298, "p90": 0.4987618997750114, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "cmh_m_s": { "median": 0.050527052036697565, "p90": 0.09166085184850266, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "rhcm_continuous_estimate_m_s": { "median": 0.18121469172088842, "p90": 0.4598460591705986, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "rhcm_discrete_m_s": { "median": 0.18115578499343343, "p90": 0.4597010420885303, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "wms": { "median": 0.7730772243427733, "p90": 0.9511009844081775, "valid_pairs": 568, "valid_duration_s": 18.933333333333337, "coverage": 0.9482470784641069 }, "rham_deg_s": { "median": 44.28191265644322, "p90": 144.67210848223255, "valid_pairs": 570, "valid_duration_s": 19.0, "coverage": 0.9515859766277127 }, "speed_spearman": 0.3499672529811289, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.031195499479043032, "p90": 0.14952179121009956, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549247 }, "cmh_m_s": { "median": 0.05543700000748993, "p90": 0.1077070633205288, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549247 }, "rhcm_continuous_estimate_m_s": { "median": 0.08282918315774052, "p90": 0.18299625107836082, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549247 }, "rhcm_discrete_m_s": { "median": 0.08285886100356143, "p90": 0.1831079659951079, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549247 }, "wms": { "median": 0.8595169081660841, "p90": 0.9758064216875447, "valid_pairs": 562, "valid_duration_s": 18.733333333333334, "coverage": 0.9382303839732887 }, "rham_deg_s": { "median": 16.61080305679053, "p90": 57.92332366002635, "valid_pairs": 574, "valid_duration_s": 19.133333333333333, "coverage": 0.9582637729549247 }, "speed_spearman": 0.01593954906086734, "hand_class": "low", "relative_class": "low", "relative_pose_class": "moving" } }, "hand_both_class": "mixed", "relative_both_class": "mixed", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06116824786105195, "camera_angular_speed_p90_deg_s": 13.966974060209177, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.4987618997750114, "hand_camera_speed_p90_m_s": 0.4597010420885303, "hand_world_valid_fraction": 0.955, "hand_camera_valid_fraction": 0.955, "cmh_p90_m_s": 0.1077070633205288, "rhcm_p90_m_s": 0.4597010420885303, "rham_p90_deg_s": 144.67210848223255, "wms_p90": 0.9758064216875447, "gt_coverage": 0.9582637729549247, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "a8c01b93592bedd21c210e104898a51223d95729613ad5f5fd488ca411e0c5a9", "metadata_sha256": "acdfbd61b92f56141b08bb50f24530bba417145b4bfb927b0adf9f089f3e26ea", "source_file": "long20_20260918/local/train/previews/fd57bc093781ef7f6d597bd1/reference.npz", "window_s": [ 0.06666666666666667, 20.033333333333335 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.578092683617719e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "dataset": "ARCTIC", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "515ab4d40dc54a33d9e042d9b38f77b75c9a5faca7b3ee3e0f6b39441024686d", "mano_sha256": "cf605cfb710a34636489898185fe7ee4ff0bc73493dcf9134a1a8f4ca51a7eff", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native ARCTIC MANO parameters; official fitted annotation", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "native source fitted MANO annotation; visualization only, not benchmark input", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "training_review", "context_id": "fe67631965976544713e8970", "sample_id": "4dd38188e2986a817d52ce9a", "file": "20_seconds/ARCTIC/4dd38188e2986a817d52ce9a_MANO_GT_20.00s.mp4", "recording_id": "s09/notebook_use_04", "source_first_frame": "90", "source_last_frame": "689", "source_split": "external_train", "training_label_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 52/600 output frames; may include genuine out-of-view frames", "id": "fe67631965976544713e8970", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/ARCTIC/4dd38188e2986a817d52ce9a_MANO_GT_20.00s.mp4", "bytes": 5001418, "width": 1346, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.02199825327178376, "p90": 0.06422807482097256, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "camera_rotation": { "median": 4.836105832041496, "p90": 9.801810153086063, "valid_pairs": 599, "valid_duration_s": 19.966666666666665, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.10918136298875451, "p90": 0.314481782432152, "valid_pairs": 547, "valid_duration_s": 18.233333333333334, "coverage": 0.9131886477462439 }, "cmh_m_s": { "median": 0.03380730308539145, "p90": 0.08068808994970728, "valid_pairs": 547, "valid_duration_s": 18.233333333333334, "coverage": 0.9131886477462439 }, "rhcm_continuous_estimate_m_s": { "median": 0.11566903562976326, "p90": 0.3430447956261781, "valid_pairs": 547, "valid_duration_s": 18.233333333333334, "coverage": 0.9131886477462439 }, "rhcm_discrete_m_s": { "median": 0.11567124921062595, "p90": 0.3430338167544717, "valid_pairs": 547, "valid_duration_s": 18.233333333333334, "coverage": 0.9131886477462439 }, "wms": { "median": 0.8159727720748514, "p90": 0.9673181061289473, "valid_pairs": 547, "valid_duration_s": 18.233333333333334, "coverage": 0.9131886477462439 }, "rham_deg_s": { "median": 48.78146993813521, "p90": 189.15945269472127, "valid_pairs": 547, "valid_duration_s": 18.233333333333334, "coverage": 0.9131886477462439 }, "speed_spearman": 0.049449596294363365, "hand_class": "high", "relative_class": "high", "relative_pose_class": "moving" }, "right": { "whm_m_s": { "median": 0.1332333834965029, "p90": 0.3234814629845048, "valid_pairs": 471, "valid_duration_s": 15.700000000000001, "coverage": 0.7863105175292154 }, "cmh_m_s": { "median": 0.03275145875396049, "p90": 0.08025993285779495, "valid_pairs": 471, "valid_duration_s": 15.700000000000001, "coverage": 0.7863105175292154 }, "rhcm_continuous_estimate_m_s": { "median": 0.12738112884931663, "p90": 0.3271161014979531, "valid_pairs": 471, "valid_duration_s": 15.700000000000001, "coverage": 0.7863105175292154 }, "rhcm_discrete_m_s": { "median": 0.1274064706950138, "p90": 0.32704443632038266, "valid_pairs": 471, "valid_duration_s": 15.700000000000001, "coverage": 0.7863105175292154 }, "wms": { "median": 0.8167749773042174, "p90": 0.9628479751071154, "valid_pairs": 470, "valid_duration_s": 15.666666666666666, "coverage": 0.7846410684474124 }, "rham_deg_s": { "median": 71.88496060963925, "p90": 168.95256465627554, "valid_pairs": 471, "valid_duration_s": 15.700000000000001, "coverage": 0.7863105175292154 }, "speed_spearman": -0.1502572577699886, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "moving", "camera_class": "low", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.06422807482097256, "camera_angular_speed_p90_deg_s": 9.801810153086063, "camera_valid_fraction": 1.0, "camera_pairs": 599, "hand_world_speed_p90_m_s": 0.3234814629845048, "hand_camera_speed_p90_m_s": 0.3430338167544717, "hand_world_valid_fraction": 0.8525, "hand_camera_valid_fraction": 0.8525, "cmh_p90_m_s": 0.08068808994970728, "rhcm_p90_m_s": 0.3430338167544717, "rham_p90_deg_s": 189.15945269472127, "wms_p90": 0.9673181061289473, "gt_coverage": 0.9131886477462439, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "published calibrated ego camera trajectory from camera mocap markers; inverse(world2ego)", "hand_source": "published fitted MANO parameters decoded with smplx; joints derived, not direct joint measurements", "reference_sha256": "19e4150434325444cb4d0b75b2bf586574ed87a5edafb972514adc9469e90504", "metadata_sha256": "5658de796997d029f0b2a934da89430cc6b3532c4d27140138b78aa746c63ac3", "source_file": "long20_20260918/local/train/previews/fe67631965976544713e8970/reference.npz", "window_s": [ 2.933333333333333, 22.9 ], "timestamps_source": "(image frame number - subject ioi_offset)/30; nominal official 30 Hz, not sensor PTS", "frame_count": 600, "common_world_residual_max_m": 4.742874840267547e-16 }, "camera_display_source": "native source camera trajectory", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方标定/mocap", "kind": "official", "detail": "使用官方相机 mocap 与标定参考轨迹。" } } }, { "id": "2f70235bf07cab120270536d", "context_id": "2f70235bf07cab120270536d", "sample_id": "2f70235bf07cab120270536d", "dataset": "Ego4D", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "a3a17e96a9fdbfe119bbac11622c28f37713f5fa357fc7eb5581a59f30ee5ed8", "mano_sha256": "c00317fb66b0d1057d689afd9f74678da7b0216a19bc69408c437ca1c24f98a7", "file": "20_seconds/Ego4D/2f70235bf07cab120270536d_P3R3_FIT_20260921_20.00s.mp4", "aliases": [], "superseded_files": [ "20_seconds/Ego4D/2f70235bf07cab120270536d_P3R3_PREDICTED_20.00s.mp4" ], "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R3 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "training_allowed": false, "video_verified": true, "recording_id": "1314208b-b64d-4ebc-bc9a-fc05c091ed45", "source_first_frame": 81, "source_last_frame": 678, "source_split": "official_FHO_val", "sequence_mapping": "verified original source frame map; exact recording ID", "pipeline_revision": "surface_fit_restore_20260921", "duration_s": 20.0, "frames": 200, "fps": 10.0, "bytes": 3949595, "width": 1282, "height": 508, "codec": "h264", "training_label_source": "所选官方 FHO 记录提供动作/框标注；未发现 MANO 参数。当前 MANO 不是官方 GT。", "quality_warnings": "单目估计 K、深度与相机轨迹；遮挡、出画、低照度及运动模糊下仍可有漏检、错位或错误手。", "camera_display_source": "ViPE RGB SLAM 估计；不是官方相机 GT", "source_original_video_sha256": "43fecfa7db4c574ad888f0f98b4612c478c3c0c5659f59e581570b65d598e577", "source_original_duration_s": 25.4, "source_start_s": 2.7, "source_scenarios": [ "Cooking" ], "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "没有经过核验的手部与相机共同世界系 GT 轨迹；当前预测不用于参考 motion 筛选。" }, "provenance": { "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "DA3 预测 RGB 深度，经估计 K 与 ViPE 轨迹融合；尺度以 P3R3 MANO 作可视化估计锚点，不是传感器 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "ViPE 从本片段 RGB 估计相机轨迹，GeoCalib 估计 K；不是官方相机位置 GT。" }, "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" } }, "mano_display_kind": "管线自造", "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/Ego4D/2f70235bf07cab120270536d_P3R3_FIT_20260921_20.00s.mp4", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "id": "879b9e1f8b6332b9c9eb63c4", "context_id": "879b9e1f8b6332b9c9eb63c4", "sample_id": "879b9e1f8b6332b9c9eb63c4", "dataset": "Ego4D", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "8c831bb06625b84dc0bcf4225bd8c35f2992514b3c7360ac622a240ff0efc9b3", "mano_sha256": "d4e9b0c2ebdbbc8489cfa0f00a3969ab1ef0df766ecf35ff343038bd6bbe6e6c", "file": "20_seconds/Ego4D/879b9e1f8b6332b9c9eb63c4_P3R3_FIT_20260921_20.00s.mp4", "aliases": [], "superseded_files": [ "20_seconds/Ego4D/879b9e1f8b6332b9c9eb63c4_P3R3_PREDICTED_20.00s.mp4" ], "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R3 predicted MANO; visual scale anchor, not GT", "manual_note": "修复后仍有腕部宽大、手形/尺寸异常；未通过训练标签质量审核。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "training_allowed": false, "video_verified": true, "recording_id": "54d4c567-1b9d-448c-a60d-d7911fe7b95d", "source_first_frame": 90, "source_last_frame": 687, "source_split": "official_FHO_val", "sequence_mapping": "verified original source frame map; exact recording ID", "pipeline_revision": "surface_fit_restore_20260921", "duration_s": 20.0, "frames": 200, "fps": 10.0, "bytes": 4055401, "width": 1282, "height": 508, "codec": "h264", "training_label_source": "所选官方 FHO 记录提供动作/框标注；未发现 MANO 参数。当前 MANO 不是官方 GT。", "quality_warnings": "单目估计 K、深度与相机轨迹；遮挡、出画、低照度及运动模糊下仍可有漏检、错位或错误手。", "camera_display_source": "ViPE RGB SLAM 估计；不是官方相机 GT", "source_original_video_sha256": "8087ef5a8ca9d5eeef9cf7bb710725460c1430fd759b1ef1aec6be3d36e7ca06", "source_original_duration_s": 26.0, "source_start_s": 3.0, "source_scenarios": [ "Scooter mechanic" ], "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "没有经过核验的手部与相机共同世界系 GT 轨迹；当前预测不用于参考 motion 筛选。" }, "provenance": { "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "DA3 预测 RGB 深度，经估计 K 与 ViPE 轨迹融合；尺度以 P3R3 MANO 作可视化估计锚点，不是传感器 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "ViPE 从本片段 RGB 估计相机轨迹，GeoCalib 估计 K；不是官方相机位置 GT。" }, "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" } }, "mano_display_kind": "管线自造", "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/Ego4D/879b9e1f8b6332b9c9eb63c4_P3R3_FIT_20260921_20.00s.mp4", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "id": "bd274c620a99c9abfcdb6100", "context_id": "bd274c620a99c9abfcdb6100", "sample_id": "bd274c620a99c9abfcdb6100", "dataset": "Ego4D", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "30807307dc1ba04c15ca707e9c489f920e57aef00ed3a1b60b8f9a77c1fd8908", "mano_sha256": "273aa79fef985a53654953c98dee29b871aab00f32ac080a75d421317baf6e71", "file": "20_seconds/Ego4D/bd274c620a99c9abfcdb6100_P3R3_FIT_20260921_20.00s.mp4", "aliases": [], "superseded_files": [ "20_seconds/Ego4D/bd274c620a99c9abfcdb6100_P3R3_PREDICTED_20.00s.mp4" ], "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R3 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "training_allowed": false, "video_verified": true, "recording_id": "0a7c1b9a-6c98-4f74-a8ad-3f06689d0146", "source_first_frame": 134, "source_last_frame": 731, "source_split": "official_FHO_val", "sequence_mapping": "verified original source frame map; exact recording ID", "pipeline_revision": "surface_fit_restore_20260921", "duration_s": 20.0, "frames": 200, "fps": 10.0, "bytes": 4916824, "width": 1282, "height": 508, "codec": "h264", "training_label_source": "所选官方 FHO 记录提供动作/框标注；未发现 MANO 参数。当前 MANO 不是官方 GT。", "quality_warnings": "单目估计 K、深度与相机轨迹；遮挡、出画、低照度及运动模糊下仍可有漏检、错位或错误手。", "camera_display_source": "ViPE RGB SLAM 估计；不是官方相机 GT", "source_original_video_sha256": "c15aa111da3670f7287a76b6e3713433ed2d603da4f009f459816c82c64378ee", "source_original_duration_s": 28.933333333333334, "source_start_s": 4.466666666666667, "source_scenarios": [ "Carpenter" ], "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "没有经过核验的手部与相机共同世界系 GT 轨迹；当前预测不用于参考 motion 筛选。" }, "provenance": { "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "DA3 预测 RGB 深度，经估计 K 与 ViPE 轨迹融合；尺度以 P3R3 MANO 作可视化估计锚点，不是传感器 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "ViPE 从本片段 RGB 估计相机轨迹，GeoCalib 估计 K；不是官方相机位置 GT。" }, "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" } }, "mano_display_kind": "管线自造", "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/Ego4D/bd274c620a99c9abfcdb6100_P3R3_FIT_20260921_20.00s.mp4", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "id": "d3a31281f9ca77537be87fe4", "context_id": "d3a31281f9ca77537be87fe4", "sample_id": "d3a31281f9ca77537be87fe4", "dataset": "Ego4D", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "4ba2e1b75d089c750f8298dd71d749e31c558d278554179f72f79dc50d6477cb", "mano_sha256": "e9f0107ff836d8a98fedf9314f8a68aab64a73e39534e272f8da21f3a9de6fab", "file": "20_seconds/Ego4D/d3a31281f9ca77537be87fe4_P3R3_FIT_20260921_20.00s.mp4", "aliases": [], "superseded_files": [ "20_seconds/Ego4D/d3a31281f9ca77537be87fe4_P3R3_PREDICTED_20.00s.mp4" ], "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R3 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "training_allowed": false, "video_verified": true, "recording_id": "6a13b88e-fedc-4577-bd21-c158bf4c180e", "source_first_frame": 325, "source_last_frame": 922, "source_split": "official_FHO_train", "sequence_mapping": "verified original source frame map; exact recording ID", "pipeline_revision": "surface_fit_restore_20260921", "duration_s": 20.0, "frames": 200, "fps": 10.0, "bytes": 2667902, "width": 1282, "height": 508, "codec": "h264", "training_label_source": "所选官方 FHO 记录提供动作/框标注；未发现 MANO 参数。当前 MANO 不是官方 GT。", "quality_warnings": "单目估计 K、深度与相机轨迹；遮挡、出画、低照度及运动模糊下仍可有漏检、错位或错误手。", "camera_display_source": "ViPE RGB SLAM 估计；不是官方相机 GT", "source_original_video_sha256": "c7f1ae7c2da14eab4dd2d194eeb92895775af62b04fe03797e71bd54fcc3f02e", "source_original_duration_s": 41.7, "source_start_s": 10.833333333333334, "source_scenarios": [ "Cleaning / laundry" ], "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "没有经过核验的手部与相机共同世界系 GT 轨迹；当前预测不用于参考 motion 筛选。" }, "provenance": { "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "DA3 预测 RGB 深度，经估计 K 与 ViPE 轨迹融合；尺度以 P3R3 MANO 作可视化估计锚点，不是传感器 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "ViPE 从本片段 RGB 估计相机轨迹，GeoCalib 估计 K；不是官方相机位置 GT。" }, "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" } }, "mano_display_kind": "管线自造", "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/Ego4D/d3a31281f9ca77537be87fe4_P3R3_FIT_20260921_20.00s.mp4", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "id": "e1d6531dfb6527de78e75799", "context_id": "e1d6531dfb6527de78e75799", "sample_id": "e1d6531dfb6527de78e75799", "dataset": "Ego4D", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "eac8d4a14a82d86799315b322b5f57ba0bcaefead084183a6a72af981f143213", "mano_sha256": "b219dafdd9ee8b9aa87c19650f0e3abaee8a9eb05df9800dc335f875543a2f6a", "file": "20_seconds/Ego4D/e1d6531dfb6527de78e75799_P3R3_FIT_20260921_20.00s.mp4", "aliases": [], "superseded_files": [ "20_seconds/Ego4D/e1d6531dfb6527de78e75799_P3R3_PREDICTED_20.00s.mp4" ], "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R3 predicted MANO; visual scale anchor, not GT", "manual_note": "开头悬浮位置已纠正；边缘及模糊帧仍有偏差，未自动审核通过。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "training_allowed": false, "video_verified": true, "recording_id": "81aaf910-e32a-46f0-a1e9-a45fe5e19e25", "source_first_frame": 71, "source_last_frame": 668, "source_split": "official_FHO_val", "sequence_mapping": "verified original source frame map; exact recording ID", "pipeline_revision": "surface_fit_restore_20260921", "duration_s": 20.0, "frames": 200, "fps": 10.0, "bytes": 2936595, "width": 1282, "height": 508, "codec": "h264", "training_label_source": "所选官方 FHO 记录提供动作/框标注；未发现 MANO 参数。当前 MANO 不是官方 GT。", "quality_warnings": "单目估计 K、深度与相机轨迹；遮挡、出画、低照度及运动模糊下仍可有漏检、错位或错误手。", "camera_display_source": "ViPE RGB SLAM 估计；不是官方相机 GT", "source_original_video_sha256": "534b0e36bff2d2395370cc56f1621a256aba0530de982752d879ca44cd69fba5", "source_original_duration_s": 24.733333333333334, "source_start_s": 2.3666666666666667, "source_scenarios": [ "Practicing a musical instrument" ], "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "没有经过核验的手部与相机共同世界系 GT 轨迹；当前预测不用于参考 motion 筛选。" }, "provenance": { "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "DA3 预测 RGB 深度，经估计 K 与 ViPE 轨迹融合；尺度以 P3R3 MANO 作可视化估计锚点，不是传感器 GT。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "ViPE 从本片段 RGB 估计相机轨迹，GeoCalib 估计 K；不是官方相机位置 GT。" }, "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" } }, "mano_display_kind": "管线自造", "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/Ego4D/e1d6531dfb6527de78e75799_P3R3_FIT_20260921_20.00s.mp4", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "c982b7fefab829d570ef252d4be2a204528b242e3fc61ac5a61e46cf2c0b68d4", "mano_sha256": "dc8d05f8e6b81728c809dd12e1c63f1ac556326b20a78ffd4bfdd83fca4722dd", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "034050584e93b92f78754c8d", "sample_id": "034050584e93b92f78754c8d", "file": "short_originals_under_20s/EgoDex/034050584e93b92f78754c8d_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/wash_fruit/0", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "034050584e93b92f78754c8d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/034050584e93b92f78754c8d_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1830201, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/97a3019caf2d7d616ac17961_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/034050584e93b92f78754c8d_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "4907c3a0f1abe65dafd13e66ae9b01106bb0764cfac5aec4894c8c58f2283401", "mano_sha256": "f2428e56207831128b099bc01770d305b9170071d58a3557be52e551daf952ca", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "097101093e9f96a8d5a4e7d1", "sample_id": "097101093e9f96a8d5a4e7d1", "file": "short_originals_under_20s/EgoDex/097101093e9f96a8d5a4e7d1_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/clean_surface/1", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "097101093e9f96a8d5a4e7d1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/097101093e9f96a8d5a4e7d1_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1541270, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/b0d36f266b1103e4385e335b_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/097101093e9f96a8d5a4e7d1_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "c6f8d2da0494c76048bda4e58afd5c98be4e1c3fe11db7184c2a041670977f58", "mano_sha256": "55fb18973a3af4e86b0dbfa787a8920941c9956e0bfe4dddd7ddd25a178ef4f5", "duration_s": 3.5, "frames": 35, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "098630fabefc333ebdfa9fa8", "sample_id": "098630fabefc333ebdfa9fa8", "file": "short_originals_under_20s/EgoDex/098630fabefc333ebdfa9fa8_P3R3_FIT_20260921_3.50s.mp4", "recording_id": "test/assemble_disassemble_legos/23", "source_first_frame": 0, "source_last_frame": 102, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "098630fabefc333ebdfa9fa8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/098630fabefc333ebdfa9fa8_P3R3_FIT_20260921_3.50s.mp4", "bytes": 607408, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/098630fabefc333ebdfa9fa8_P3R2_PREDICTED_3.50s.mp4", "short_originals_under_20s/EgoDex/098630fabefc333ebdfa9fa8_P3R3_PREDICTED_3.50s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "b3be4c6c272b1538941e85ff540adefd9c8a4a4898e3cd1fce169e40e7bf6003", "mano_sha256": "31f67ce860028b89b7c15e708ab8c1165324f0376c0a0e314f364ae03c2aec3e", "duration_s": 7.3, "frames": 73, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "09bd02abb3a2af8f3dcee8ba", "sample_id": "09bd02abb3a2af8f3dcee8ba", "file": "short_originals_under_20s/EgoDex/09bd02abb3a2af8f3dcee8ba_P3R3_FIT_20260921_7.30s.mp4", "recording_id": "test/basic_pick_place/100", "source_first_frame": 0, "source_last_frame": 216, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "09bd02abb3a2af8f3dcee8ba", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/09bd02abb3a2af8f3dcee8ba_P3R3_FIT_20260921_7.30s.mp4", "bytes": 1229112, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/60878f6215672e6a8b32d234_P3R2_PREDICTED_7.30s.mp4", "short_originals_under_20s/EgoDex/09bd02abb3a2af8f3dcee8ba_P3R3_PREDICTED_7.30s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "158f5c94b8f6ddb202cfe1dbef4be814010d4cd89ce487e3d7c1410aa4128c02", "mano_sha256": "7f3f6694e10bb679cca70ddd3a68e2f93798138ac54598c42f6d9318848dcb32", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "0d64a11d465c82a29fda0775", "sample_id": "0d64a11d465c82a29fda0775", "file": "20_seconds/EgoDex/0d64a11d465c82a29fda0775_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "test/clip_unclip_papers/3", "source_first_frame": 423, "source_last_frame": 1020, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "0d64a11d465c82a29fda0775", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EgoDex/0d64a11d465c82a29fda0775_P3R3_FIT_20260921_20.00s.mp4", "bytes": 3128936, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "20_seconds/EgoDex/c0d2ffe3c6c7f353b997cb3c_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EgoDex/0d64a11d465c82a29fda0775_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "868522a715bb76ff95ab1e0a5adacf2f09fb1f73c1fec1334c2276f09c9da5ff", "mano_sha256": "22bcd6b0c3de3e4f30ce55f389b4e2cfc6f5678a5672a03c28af7524614ef07b", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "0f28b1c23518a1fa6a956fab", "sample_id": "0f28b1c23518a1fa6a956fab", "file": "short_originals_under_20s/EgoDex/0f28b1c23518a1fa6a956fab_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/knead_slime/18", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "0f28b1c23518a1fa6a956fab", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/0f28b1c23518a1fa6a956fab_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1546886, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/b2dc0239ec6ef5cb1f644bbe_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/0f28b1c23518a1fa6a956fab_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "caeec6f9a04c3de79710ef5ce027ca61318a9d9aa844a62bc4e304047dc98433", "mano_sha256": "cafa342dee439e22c61b572dbfe71d26945d7971d1b25934a223dfbba441d0a3", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "1a56ecd23ec27cf70c5daad8", "sample_id": "1a56ecd23ec27cf70c5daad8", "file": "20_seconds/EgoDex/1a56ecd23ec27cf70c5daad8_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "test/fidget_magnetic_spinner_rings/7", "source_first_frame": 1412, "source_last_frame": 2009, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "1a56ecd23ec27cf70c5daad8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EgoDex/1a56ecd23ec27cf70c5daad8_P3R3_FIT_20260921_20.00s.mp4", "bytes": 2274101, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "20_seconds/EgoDex/7ee0583a8aa8393d8978fa46_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EgoDex/1a56ecd23ec27cf70c5daad8_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "c545c43921ddc192c6a330804325211332fadac6670c70d314636c1843912592", "mano_sha256": "70a23270753aa07e4cf7e23f00b7d405fc627698464506e7ac22df03c5832214", "duration_s": 3.0, "frames": 30, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "1be9c315d5a0b4269b80d3c9", "sample_id": "1be9c315d5a0b4269b80d3c9", "file": "short_originals_under_20s/EgoDex/1be9c315d5a0b4269b80d3c9_P3R3_FIT_20260921_3.00s.mp4", "recording_id": "test/stack_unstack_plates/50", "source_first_frame": 0, "source_last_frame": 87, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "1be9c315d5a0b4269b80d3c9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/1be9c315d5a0b4269b80d3c9_P3R3_FIT_20260921_3.00s.mp4", "bytes": 476379, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/1be9c315d5a0b4269b80d3c9_P3R2_PREDICTED_3.00s.mp4", "short_originals_under_20s/EgoDex/1be9c315d5a0b4269b80d3c9_P3R3_PREDICTED_3.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "093222d8c67fa682b124a421300c5eef2026d0b1ef2707f7b300b8d672137cd1", "mano_sha256": "70c1d0235942de36aa6cdb0aa74a55de16f297880ad4ebae27d33ad3bd419adc", "duration_s": 8.5, "frames": 85, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "208fb7bfb38be5bd009e8967", "sample_id": "208fb7bfb38be5bd009e8967", "file": "short_originals_under_20s/EgoDex/208fb7bfb38be5bd009e8967_P3R3_FIT_20260921_8.50s.mp4", "recording_id": "test/insert_remove_airpods/41", "source_first_frame": 0, "source_last_frame": 252, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "208fb7bfb38be5bd009e8967", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/208fb7bfb38be5bd009e8967_P3R3_FIT_20260921_8.50s.mp4", "bytes": 1424139, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/c0a1e20fa462d0b786c19f70_P3R2_PREDICTED_8.50s.mp4", "short_originals_under_20s/EgoDex/208fb7bfb38be5bd009e8967_P3R3_PREDICTED_8.50s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "d9ea441fe6e52612083d886d71294df996e15982e80c8db360f8abbe44349667", "mano_sha256": "288e451adce341ccb93be576e685623cade8bde97c58cd4087219a851b2f4189", "duration_s": 2.4, "frames": 24, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "24592b62a8e6863f83d46259", "sample_id": "24592b62a8e6863f83d46259", "file": "short_originals_under_20s/EgoDex/24592b62a8e6863f83d46259_P3R3_FIT_20260921_2.40s.mp4", "recording_id": "test/charge_uncharge_airpods/7", "source_first_frame": 0, "source_last_frame": 69, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "24592b62a8e6863f83d46259", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/24592b62a8e6863f83d46259_P3R3_FIT_20260921_2.40s.mp4", "bytes": 403671, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/24592b62a8e6863f83d46259_P3R2_PREDICTED_2.40s.mp4", "short_originals_under_20s/EgoDex/24592b62a8e6863f83d46259_P3R3_PREDICTED_2.40s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "dcbef2fc9c9f5995b67b5ee5d74c53a1f8179c377b2d09a3ab80478741ace20e", "mano_sha256": "c1d0a62a76918e49968979aff9221d26faeae555e2649d5aef447fae759d774d", "duration_s": 3.4, "frames": 34, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "25d0ead41e702d6a38d0fb25", "sample_id": "25d0ead41e702d6a38d0fb25", "file": "short_originals_under_20s/EgoDex/25d0ead41e702d6a38d0fb25_P3R3_FIT_20260921_3.40s.mp4", "recording_id": "test/arrange_topple_dominoes/3", "source_first_frame": 0, "source_last_frame": 99, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "25d0ead41e702d6a38d0fb25", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/25d0ead41e702d6a38d0fb25_P3R3_FIT_20260921_3.40s.mp4", "bytes": 593227, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/25d0ead41e702d6a38d0fb25_P3R2_PREDICTED_3.40s.mp4", "short_originals_under_20s/EgoDex/25d0ead41e702d6a38d0fb25_P3R3_PREDICTED_3.40s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "6f282bccc7e7f3d0b7020a17a0a14cd370e0b89cbb966e55beaa082b59d0102c", "mano_sha256": "36ede4e6e9992c1ff9791f8e7b1274a97bb416d5f1cb9aa3186159a497fb0738", "duration_s": 4.3, "frames": 43, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "268f4348f9c636fb35e5edac", "sample_id": "268f4348f9c636fb35e5edac", "file": "short_originals_under_20s/EgoDex/268f4348f9c636fb35e5edac_P3R3_FIT_20260921_4.30s.mp4", "recording_id": "test/assemble_disassemble_furniture_bench_desk/24", "source_first_frame": 0, "source_last_frame": 126, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "268f4348f9c636fb35e5edac", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/268f4348f9c636fb35e5edac_P3R3_FIT_20260921_4.30s.mp4", "bytes": 694434, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/268f4348f9c636fb35e5edac_P3R2_PREDICTED_4.30s.mp4", "short_originals_under_20s/EgoDex/268f4348f9c636fb35e5edac_P3R3_PREDICTED_4.30s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "75db50e6ea148769d3d6079f2c5e7523fcb184bc451b7c460c01a4b9803524ed", "mano_sha256": "35c23db5c359f78b379ffd8d03f819cefab52ee508e7d46928d437c70c5d73fd", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "2afb94deb24512ac27ca2244", "sample_id": "2afb94deb24512ac27ca2244", "file": "short_originals_under_20s/EgoDex/2afb94deb24512ac27ca2244_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/sweep_dustpan/3", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "2afb94deb24512ac27ca2244", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/2afb94deb24512ac27ca2244_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1899343, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/644062d69af4245c13e79044_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/2afb94deb24512ac27ca2244_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "a7b89913e485ad48e6ff768157e0408dad5c4b5691d40ca74668c3e66c4aa03c", "mano_sha256": "3ecc155e2451dbe1f763b22b8f54355978e0cdd0ca7607ec9a970a4d712e5994", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "2bc3b5c8a6c996dee7761f2f", "sample_id": "2bc3b5c8a6c996dee7761f2f", "file": "short_originals_under_20s/EgoDex/2bc3b5c8a6c996dee7761f2f_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/assemble_jenga/3", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "2bc3b5c8a6c996dee7761f2f", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/2bc3b5c8a6c996dee7761f2f_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1582732, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/4798e7b17c9cc453c7e3fac1_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/2bc3b5c8a6c996dee7761f2f_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "251d0af96e0f96cd77771d8540ad35d9d54c126dfb4790bec860f08d1d9c932b", "mano_sha256": "51a657b01297ce9d1670551b49cf1e487bef2e3adf4972d23e92dbb74e1f4001", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "2c2ae2c3a6fd3f1f4b12c7bb", "sample_id": "2c2ae2c3a6fd3f1f4b12c7bb", "file": "short_originals_under_20s/EgoDex/2c2ae2c3a6fd3f1f4b12c7bb_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/fold_unfold_paper_origami/7", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "2c2ae2c3a6fd3f1f4b12c7bb", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/2c2ae2c3a6fd3f1f4b12c7bb_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1498035, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/ed2029f18d3d78d9daa3f67e_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/2c2ae2c3a6fd3f1f4b12c7bb_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "e3b88040c735d8868c4a83d8489ad61de6bd03e43792dcfd1bd91cef380178e5", "mano_sha256": "f2ede610a6613bb81701fda442fd186a364ae401defdc90dfb9be58f87c362a1", "duration_s": 8.2, "frames": 82, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "2dcc732c3279a298acaeda92", "sample_id": "2dcc732c3279a298acaeda92", "file": "short_originals_under_20s/EgoDex/2dcc732c3279a298acaeda92_P3R3_FIT_20260921_8.20s.mp4", "recording_id": "test/stock_unstock_fridge/18", "source_first_frame": 0, "source_last_frame": 243, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "no valid P3R2_PREDICTED mesh in 1/82 output frames; may include genuine out-of-view frames", "id": "2dcc732c3279a298acaeda92", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/2dcc732c3279a298acaeda92_P3R3_FIT_20260921_8.20s.mp4", "bytes": 1470424, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/1e88a58d7bfd54740a2f8baa_P3R2_PREDICTED_8.20s.mp4", "short_originals_under_20s/EgoDex/2dcc732c3279a298acaeda92_P3R3_PREDICTED_8.20s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "1d97633bee4f1c2e54f58ec344f2c452a8e2265bfbaed081dffc3b817a066512", "mano_sha256": "fb069e35ed7865363f2c878ed042d9b1b154b878c57d6722c512dd35e82b53fa", "duration_s": 2.4, "frames": 24, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "3b0b5a1ea0ba11cf002a6d56", "sample_id": "3b0b5a1ea0ba11cf002a6d56", "file": "short_originals_under_20s/EgoDex/3b0b5a1ea0ba11cf002a6d56_P3R3_FIT_20260921_2.40s.mp4", "recording_id": "test/pour/6", "source_first_frame": 0, "source_last_frame": 69, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "3b0b5a1ea0ba11cf002a6d56", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/3b0b5a1ea0ba11cf002a6d56_P3R3_FIT_20260921_2.40s.mp4", "bytes": 409251, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/3b0b5a1ea0ba11cf002a6d56_P3R2_PREDICTED_2.40s.mp4", "short_originals_under_20s/EgoDex/3b0b5a1ea0ba11cf002a6d56_P3R3_PREDICTED_2.40s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "5a775d7bf51f19a5b2b37e2837313047dcfe7d723ace5b3d93753fdb9f304eaa", "mano_sha256": "55b0e9776f626c3c89b33623d6e85fdf4b0f56727f4fe57432a11b3b47baf28c", "duration_s": 4.3, "frames": 43, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "3eae5866c303741faa2b7b11", "sample_id": "3eae5866c303741faa2b7b11", "file": "short_originals_under_20s/EgoDex/3eae5866c303741faa2b7b11_P3R3_FIT_20260921_4.30s.mp4", "recording_id": "test/insert_remove_cups_from_rack/6", "source_first_frame": 0, "source_last_frame": 126, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "3eae5866c303741faa2b7b11", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/3eae5866c303741faa2b7b11_P3R3_FIT_20260921_4.30s.mp4", "bytes": 894732, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/3eae5866c303741faa2b7b11_P3R2_PREDICTED_4.30s.mp4", "short_originals_under_20s/EgoDex/3eae5866c303741faa2b7b11_P3R3_PREDICTED_4.30s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "f8143813a2ef4ece07bd97437741da4667d56e85fa2f2c7ac8d32cce7dfdabe8", "mano_sha256": "2b10a9f2cc05852e9f16b052af98b4a9fceae32e389a26817485cfc7fa589495", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "402cd8c39ce2fbad52005d48", "sample_id": "402cd8c39ce2fbad52005d48", "file": "20_seconds/EgoDex/402cd8c39ce2fbad52005d48_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "test/put_toothpaste_on_toothbrush/6", "source_first_frame": 2, "source_last_frame": 599, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "402cd8c39ce2fbad52005d48", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EgoDex/402cd8c39ce2fbad52005d48_P3R3_FIT_20260921_20.00s.mp4", "bytes": 2657641, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "20_seconds/EgoDex/cdf2c35a127672cbdd6f0819_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EgoDex/402cd8c39ce2fbad52005d48_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "ae84deb20469a7176aa5bfde9fd35b3fde10e6ed736d714b47392adeec71077d", "mano_sha256": "a62bf5405bd7b463e5775359952d76b63d85ca0be0b231f2c22d39a68e2b1d0e", "duration_s": 4.1, "frames": 41, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "59e752843a4fb7f162465775", "sample_id": "59e752843a4fb7f162465775", "file": "short_originals_under_20s/EgoDex/59e752843a4fb7f162465775_P3R3_FIT_20260921_4.10s.mp4", "recording_id": "test/insert_remove_drawer/12", "source_first_frame": 0, "source_last_frame": 120, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "59e752843a4fb7f162465775", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/59e752843a4fb7f162465775_P3R3_FIT_20260921_4.10s.mp4", "bytes": 570808, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/59e752843a4fb7f162465775_P3R2_PREDICTED_4.10s.mp4", "short_originals_under_20s/EgoDex/59e752843a4fb7f162465775_P3R3_PREDICTED_4.10s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "0b15296d54d136d26cb959fc69ffe43818a2f7b93ae5536b0bf3d59020c2bd33", "mano_sha256": "0197c11c78fe0a1520e4c48c422bb7102f495cc0f8f39cca42b79dc67371d588", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "5bd348a7b526ed2ed6769c8e", "sample_id": "5bd348a7b526ed2ed6769c8e", "file": "20_seconds/EgoDex/5bd348a7b526ed2ed6769c8e_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "test/insert_remove_tennis_ball/1", "source_first_frame": 2, "source_last_frame": 599, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "hand_jump_over_25cm_per_output_frame", "id": "5bd348a7b526ed2ed6769c8e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EgoDex/5bd348a7b526ed2ed6769c8e_P3R3_FIT_20260921_20.00s.mp4", "bytes": 3395865, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "20_seconds/EgoDex/28daca25f3f745928b3f2016_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EgoDex/5bd348a7b526ed2ed6769c8e_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "3e4197ae704dc82c13d87857c5fcf59daa3789b07949cd2aeb76184b0d77bd25", "mano_sha256": "ae7998771b7881549937151a521003bc407444a3f94b166528f68d6bf5778b17", "duration_s": 16.0, "frames": 160, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "656dac679ae2a71f4a3c7ea0", "sample_id": "656dac679ae2a71f4a3c7ea0", "file": "short_originals_under_20s/EgoDex/656dac679ae2a71f4a3c7ea0_P3R3_FIT_20260921_16.00s.mp4", "recording_id": "test/crumple_flatten_paper/2", "source_first_frame": 0, "source_last_frame": 477, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "656dac679ae2a71f4a3c7ea0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/656dac679ae2a71f4a3c7ea0_P3R3_FIT_20260921_16.00s.mp4", "bytes": 2541531, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/21b01e5beda21a3200b8918c_P3R2_PREDICTED_16.00s.mp4", "short_originals_under_20s/EgoDex/656dac679ae2a71f4a3c7ea0_P3R3_PREDICTED_16.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "fcc038d1525f312eb3ec5e5b8fc58ad292fcfe14bf19d82b5539caa3539a6c76", "mano_sha256": "128444668ccbc54786b9abfab5eacb7611f356e44bbd0d2683fb3f32124abda1", "duration_s": 11.6, "frames": 116, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "6b7986be18679df4a2b75dde", "sample_id": "6b7986be18679df4a2b75dde", "file": "short_originals_under_20s/EgoDex/6b7986be18679df4a2b75dde_P3R3_FIT_20260921_11.60s.mp4", "recording_id": "test/wrap_unwrap_food/0", "source_first_frame": 0, "source_last_frame": 345, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "6b7986be18679df4a2b75dde", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/6b7986be18679df4a2b75dde_P3R3_FIT_20260921_11.60s.mp4", "bytes": 2301379, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/0037a5f44b26878286c8af1d_P3R2_PREDICTED_11.60s.mp4", "short_originals_under_20s/EgoDex/6b7986be18679df4a2b75dde_P3R3_PREDICTED_11.60s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "c98522a2db2b33550c10d1f3cfc4b459a7536d047a02dd0fdbef8f1bcc9a0877", "mano_sha256": "aefd4adf4dc9365d919aa37919afce5abda2dd11c3c4d3ff01812a656d0b208d", "duration_s": 2.1, "frames": 21, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "74c52055dcfef02b7739689a", "sample_id": "74c52055dcfef02b7739689a", "file": "short_originals_under_20s/EgoDex/74c52055dcfef02b7739689a_P3R3_FIT_20260921_2.10s.mp4", "recording_id": "test/insert_remove_furniture_bench_cabinet/71", "source_first_frame": 0, "source_last_frame": 60, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "74c52055dcfef02b7739689a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/74c52055dcfef02b7739689a_P3R3_FIT_20260921_2.10s.mp4", "bytes": 373311, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/74c52055dcfef02b7739689a_P3R2_PREDICTED_2.10s.mp4", "short_originals_under_20s/EgoDex/74c52055dcfef02b7739689a_P3R3_PREDICTED_2.10s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "d8eb76149b15fc8030679a2a77e08bc5fdafff6d1bc399689b2d665e58824a24", "mano_sha256": "398b88a3e70e5dd48722749cdd6a06a30b7e768132d7f1da06de139d277a3df7", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "756d8ac76c7fd0b24084b2a9", "sample_id": "756d8ac76c7fd0b24084b2a9", "file": "short_originals_under_20s/EgoDex/756d8ac76c7fd0b24084b2a9_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/clean_tableware/11", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "756d8ac76c7fd0b24084b2a9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/756d8ac76c7fd0b24084b2a9_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1964449, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/1e15fbe6e552057d1ea3167c_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/756d8ac76c7fd0b24084b2a9_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "e899e12d544ae05578e310d2bd4a202310d3c8ca5ca9517d5b9fcb0251a9bdff", "mano_sha256": "0a0f8223651fc57e54eca818f720996697f666dc00f137738c15c01fc939b24c", "duration_s": 3.5, "frames": 35, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "80e638370142d10eee3a9516", "sample_id": "80e638370142d10eee3a9516", "file": "short_originals_under_20s/EgoDex/80e638370142d10eee3a9516_P3R3_FIT_20260921_3.50s.mp4", "recording_id": "test/open_close_insert_remove_tupperware/10", "source_first_frame": 0, "source_last_frame": 102, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "80e638370142d10eee3a9516", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/80e638370142d10eee3a9516_P3R3_FIT_20260921_3.50s.mp4", "bytes": 635537, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/80e638370142d10eee3a9516_P3R2_PREDICTED_3.50s.mp4", "short_originals_under_20s/EgoDex/80e638370142d10eee3a9516_P3R3_PREDICTED_3.50s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "036e519319422ef63ce02ee3ec79cd0770ec705d22d83e9bf6d5331d68c57b1f", "mano_sha256": "c70ec67a8977d048ee83718d12e0fc9d8cf06c38a0742143b886ae26dedb03c5", "duration_s": 2.8, "frames": 28, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "83fa9eaf353dd619f368e419", "sample_id": "83fa9eaf353dd619f368e419", "file": "short_originals_under_20s/EgoDex/83fa9eaf353dd619f368e419_P3R3_FIT_20260921_2.80s.mp4", "recording_id": "test/stack/14", "source_first_frame": 0, "source_last_frame": 81, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "83fa9eaf353dd619f368e419", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/83fa9eaf353dd619f368e419_P3R3_FIT_20260921_2.80s.mp4", "bytes": 402392, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/83fa9eaf353dd619f368e419_P3R2_PREDICTED_2.80s.mp4", "short_originals_under_20s/EgoDex/83fa9eaf353dd619f368e419_P3R3_PREDICTED_2.80s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "d1716096e0172703891aa49f2da9142f4910b623a0c585ef7b2be327f9cc3928", "mano_sha256": "fe621844af7355a568bbc22235ad8237845f9277cfb8b73bcad6a2374d933df7", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "87dbf841bc3acf540bf11040", "sample_id": "87dbf841bc3acf540bf11040", "file": "20_seconds/EgoDex/87dbf841bc3acf540bf11040_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "test/slot_batteries/2", "source_first_frame": 0, "source_last_frame": 597, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "87dbf841bc3acf540bf11040", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EgoDex/87dbf841bc3acf540bf11040_P3R3_FIT_20260921_20.00s.mp4", "bytes": 3597740, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "20_seconds/EgoDex/738679c3cd9cbd8677e73599_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EgoDex/87dbf841bc3acf540bf11040_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "aec1f57275ea55bcffe47f4e5af938d3535015bdc0bc5ec26788c8640a182d00", "mano_sha256": "279ae7341a5d557a5e24f756e002c9dc6a351f6c304469b3c8a214500f0359b2", "duration_s": 7.0, "frames": 70, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "880754b3c92a2dfb43bdcff1", "sample_id": "880754b3c92a2dfb43bdcff1", "file": "short_originals_under_20s/EgoDex/880754b3c92a2dfb43bdcff1_P3R3_FIT_20260921_7.00s.mp4", "recording_id": "test/tie_untie_rubberband/7", "source_first_frame": 0, "source_last_frame": 207, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "880754b3c92a2dfb43bdcff1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/880754b3c92a2dfb43bdcff1_P3R3_FIT_20260921_7.00s.mp4", "bytes": 1186552, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/5930fcad5e26e7aa188fb51d_P3R2_PREDICTED_7.00s.mp4", "short_originals_under_20s/EgoDex/880754b3c92a2dfb43bdcff1_P3R3_PREDICTED_7.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "cb808b72bc07b68df35daa57a1586d56eb0d679406d90b4faa2e766242ad0576", "mano_sha256": "f863f7874fc1d64cee6362fdae276881308a47c405934b82c65ef1a7befe66e7", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "90dcb31b49e031c4649c69d9", "sample_id": "90dcb31b49e031c4649c69d9", "file": "short_originals_under_20s/EgoDex/90dcb31b49e031c4649c69d9_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/paint_clean_brush/16", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "90dcb31b49e031c4649c69d9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/90dcb31b49e031c4649c69d9_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1652274, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/cab93a44e7b6e35b3e9c5101_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/90dcb31b49e031c4649c69d9_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "b2191045349468aaf318bae82c6990a430094a10cb740f02da36116676a651dc", "mano_sha256": "dc57fa115049a04a1dbe00c3fd3a75022b3e5e31df7d5c0e0095227b1bd3f0aa", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "95f93ca3dacc58b9b761d59f", "sample_id": "95f93ca3dacc58b9b761d59f", "file": "short_originals_under_20s/EgoDex/95f93ca3dacc58b9b761d59f_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/sort_beads/7", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "95f93ca3dacc58b9b761d59f", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/95f93ca3dacc58b9b761d59f_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1829627, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/ce60fd25148f4dc4271cfb80_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/95f93ca3dacc58b9b761d59f_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "e1ea73a5ce9ad8b0e57e0e6c3ad363746c06d799a3c79d070ac3dc00e3b6c7fe", "mano_sha256": "90580d6a1a3314ccf74fa762648876c3350f6616312aa4165e73631cd5d4e936", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "96495b27b3e4af0cca0db848", "sample_id": "96495b27b3e4af0cca0db848", "file": "20_seconds/EgoDex/96495b27b3e4af0cca0db848_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "test/set_up_clean_up_chessboard/2", "source_first_frame": 2485, "source_last_frame": 3082, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "96495b27b3e4af0cca0db848", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EgoDex/96495b27b3e4af0cca0db848_P3R3_FIT_20260921_20.00s.mp4", "bytes": 3959208, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "20_seconds/EgoDex/8a32081ae30342d6042a00fa_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EgoDex/96495b27b3e4af0cca0db848_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "3e0931d6215a68ccc2f9e4cc2e68e75157a2486e0e260e4d2edbee967ac8f67c", "mano_sha256": "e388dcba7a12ef2a16e8ba5060b963a6952480a1dd44f0a58fdbf63440aeda5d", "duration_s": 16.2, "frames": 162, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "98b0e8088c18a98cc87411b2", "sample_id": "98b0e8088c18a98cc87411b2", "file": "short_originals_under_20s/EgoDex/98b0e8088c18a98cc87411b2_P3R3_FIT_20260921_16.20s.mp4", "recording_id": "test/assemble_disassemble_tiles/3", "source_first_frame": 0, "source_last_frame": 483, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "hand_jump_over_25cm_per_output_frame", "id": "98b0e8088c18a98cc87411b2", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/98b0e8088c18a98cc87411b2_P3R3_FIT_20260921_16.20s.mp4", "bytes": 3072715, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/3e5f11951b460a979af342c9_P3R2_PREDICTED_16.20s.mp4", "short_originals_under_20s/EgoDex/98b0e8088c18a98cc87411b2_P3R3_PREDICTED_16.20s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "4c4d49b37915477d6167b20a2c40f7fad41edd6d6a6ccc245fd2c47a025a206a", "mano_sha256": "c52391d2efab6a2bd5e66f9a31cf5215362a68a4110f5c0ff6dbc6ef4edc4e29", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "a09cd7de6f0f9553d9fcd89b", "sample_id": "a09cd7de6f0f9553d9fcd89b", "file": "short_originals_under_20s/EgoDex/a09cd7de6f0f9553d9fcd89b_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/throw_and_catch_ball/19", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "a09cd7de6f0f9553d9fcd89b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/a09cd7de6f0f9553d9fcd89b_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1534953, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/d04592e519ce5db605603581_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/a09cd7de6f0f9553d9fcd89b_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "cb9802ac4db5bc6cf6823a4d7d070fb5900d5dc0f2cc0d1a20cf66e399c0d1eb", "mano_sha256": "a8bd0c6defaa3692a23511db1d4958aabdd67e503a7880cb1fa566d3f1729bca", "duration_s": 2.4, "frames": 24, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "a1df2cc35faced1f1dd67cc3", "sample_id": "a1df2cc35faced1f1dd67cc3", "file": "short_originals_under_20s/EgoDex/a1df2cc35faced1f1dd67cc3_P3R3_FIT_20260921_2.40s.mp4", "recording_id": "test/assemble_disassemble_furniture_bench_drawer/9", "source_first_frame": 0, "source_last_frame": 69, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "a1df2cc35faced1f1dd67cc3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/a1df2cc35faced1f1dd67cc3_P3R3_FIT_20260921_2.40s.mp4", "bytes": 446505, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/a1df2cc35faced1f1dd67cc3_P3R2_PREDICTED_2.40s.mp4", "short_originals_under_20s/EgoDex/a1df2cc35faced1f1dd67cc3_P3R3_PREDICTED_2.40s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "d8591be305930b5e5855d4a26454991d104695b0759a1a4cfbc21ce0d06e9f64", "mano_sha256": "d1b4c9e70a4ef828667956079ac7e271b113e9f029c355e1b45c326ca3b18dcd", "duration_s": 10.0, "frames": 100, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "ab14fece7c649de2f5cfcc9a", "sample_id": "ab14fece7c649de2f5cfcc9a", "file": "short_originals_under_20s/EgoDex/ab14fece7c649de2f5cfcc9a_P3R3_FIT_20260921_10.00s.mp4", "recording_id": "test/point_and_click_remote/16", "source_first_frame": 0, "source_last_frame": 297, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "ab14fece7c649de2f5cfcc9a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/ab14fece7c649de2f5cfcc9a_P3R3_FIT_20260921_10.00s.mp4", "bytes": 1644567, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/50c3833208a36b6f8d1e7764_P3R2_PREDICTED_10.00s.mp4", "short_originals_under_20s/EgoDex/ab14fece7c649de2f5cfcc9a_P3R3_PREDICTED_10.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "c5d6dbd9188714e65ce80ad3b47afc15a6fec02aeb28f7e9715fbe3581271469", "mano_sha256": "3ad2bf480f97d3b6827f03597a2b9211bc008db0ccb84d0bd1c3919d2b706255", "duration_s": 11.0, "frames": 110, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "ab723b8b8a60be4d9fd89728", "sample_id": "ab723b8b8a60be4d9fd89728", "file": "short_originals_under_20s/EgoDex/ab723b8b8a60be4d9fd89728_P3R3_FIT_20260921_11.00s.mp4", "recording_id": "test/build_unstack_lego/10", "source_first_frame": 0, "source_last_frame": 327, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "ab723b8b8a60be4d9fd89728", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/ab723b8b8a60be4d9fd89728_P3R3_FIT_20260921_11.00s.mp4", "bytes": 1834428, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/52e88f40005efab02f7bad0f_P3R2_PREDICTED_11.00s.mp4", "short_originals_under_20s/EgoDex/ab723b8b8a60be4d9fd89728_P3R3_PREDICTED_11.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "c7d222898327206d42c299bf45f67cf9734620ae5c5252d341e0485ded913607", "mano_sha256": "02a11a1c6759132710551d25325aba36e24ddb05f6a5fc04467f53c2d8820745", "duration_s": 10.2, "frames": 102, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "c1ddd870ad9f64c607c2611d", "sample_id": "c1ddd870ad9f64c607c2611d", "file": "short_originals_under_20s/EgoDex/c1ddd870ad9f64c607c2611d_P3R3_FIT_20260921_10.20s.mp4", "recording_id": "test/wash_put_away_dishes/0", "source_first_frame": 0, "source_last_frame": 303, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "no valid P3R2_PREDICTED mesh in 3/102 output frames; may include genuine out-of-view frames", "id": "c1ddd870ad9f64c607c2611d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/c1ddd870ad9f64c607c2611d_P3R3_FIT_20260921_10.20s.mp4", "bytes": 1970010, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/5ac6d91c4a3996d407cbd56d_P3R2_PREDICTED_10.20s.mp4", "short_originals_under_20s/EgoDex/c1ddd870ad9f64c607c2611d_P3R3_PREDICTED_10.20s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "f971245092093cacafd75e56cf1588a16c5e7f06c1ed22bcd6c3d6918f2d148c", "mano_sha256": "cb7ca6bd05515ba5ec4f80d6a2cca78920b036712d72cb488d4d03194e5dc982", "duration_s": 13.0, "frames": 130, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "c99611d4a6cb7f2424423357", "sample_id": "c99611d4a6cb7f2424423357", "file": "short_originals_under_20s/EgoDex/c99611d4a6cb7f2424423357_P3R3_FIT_20260921_13.00s.mp4", "recording_id": "test/screw_unscrew_fingers_fixture/10", "source_first_frame": 0, "source_last_frame": 387, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "c99611d4a6cb7f2424423357", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/c99611d4a6cb7f2424423357_P3R3_FIT_20260921_13.00s.mp4", "bytes": 1923018, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/f6c373b7a6048fb6bdfcdc5d_P3R2_PREDICTED_13.00s.mp4", "short_originals_under_20s/EgoDex/c99611d4a6cb7f2424423357_P3R3_PREDICTED_13.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "304ebfd213bda60805443795b80b0239f045796c8f24a83e122a6c12156514c6", "mano_sha256": "40e5ca6f1c17ad72a7523da3b65888e30321065aa043dd4eebc920bca2eb43b7", "duration_s": 7.5, "frames": 75, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "cbe182eb3873f13bdadd6e47", "sample_id": "cbe182eb3873f13bdadd6e47", "file": "short_originals_under_20s/EgoDex/cbe182eb3873f13bdadd6e47_P3R3_FIT_20260921_7.50s.mp4", "recording_id": "test/stack_remove_jenga/9", "source_first_frame": 0, "source_last_frame": 222, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "cbe182eb3873f13bdadd6e47", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/cbe182eb3873f13bdadd6e47_P3R3_FIT_20260921_7.50s.mp4", "bytes": 1249295, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/d376a251ca798a710930df72_P3R2_PREDICTED_7.50s.mp4", "short_originals_under_20s/EgoDex/cbe182eb3873f13bdadd6e47_P3R3_PREDICTED_7.50s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "36a387a4c095ee0f527e58ead46c5cecee32047dbb836c04217f03e2cd596103", "mano_sha256": "b6e5e3737b6005836b13c221aca87aac5f5a1471148686b84bf4ab69367894df", "duration_s": 4.6, "frames": 46, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "cdb0a905c9268c04aa9845b9", "sample_id": "cdb0a905c9268c04aa9845b9", "file": "short_originals_under_20s/EgoDex/cdb0a905c9268c04aa9845b9_P3R3_FIT_20260921_4.60s.mp4", "recording_id": "test/assemble_disassemble_furniture_bench_chair/13", "source_first_frame": 0, "source_last_frame": 135, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "cdb0a905c9268c04aa9845b9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/cdb0a905c9268c04aa9845b9_P3R3_FIT_20260921_4.60s.mp4", "bytes": 659005, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/cdb0a905c9268c04aa9845b9_P3R2_PREDICTED_4.60s.mp4", "short_originals_under_20s/EgoDex/cdb0a905c9268c04aa9845b9_P3R3_PREDICTED_4.60s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "73c5a6366993efde47da9a9618164dac9602faf02bbf0b47d4906ee0b55614ca", "mano_sha256": "a4bf4b45fc76d7013b7b4af54e4f6933967fb7114d227718ed62bca8c2e97391", "duration_s": 13.7, "frames": 137, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "cf2af02dc0c82e6b2d1089f0", "sample_id": "cf2af02dc0c82e6b2d1089f0", "file": "short_originals_under_20s/EgoDex/cf2af02dc0c82e6b2d1089f0_P3R3_FIT_20260921_13.70s.mp4", "recording_id": "test/insert_remove_utensils/8", "source_first_frame": 0, "source_last_frame": 408, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "cf2af02dc0c82e6b2d1089f0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/cf2af02dc0c82e6b2d1089f0_P3R3_FIT_20260921_13.70s.mp4", "bytes": 2539085, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/dd45864e4ada80f54f986dcd_P3R2_PREDICTED_13.70s.mp4", "short_originals_under_20s/EgoDex/cf2af02dc0c82e6b2d1089f0_P3R3_PREDICTED_13.70s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "8a155e59729600655c43855fe00eece969e2850cdad0474e67cb1b66c3fafa86", "mano_sha256": "207b8d162462a6446b854f7f22a317e1abe67f571f0fd87926984a740b8df03c", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "db1f34a08fbece0550db9006", "sample_id": "db1f34a08fbece0550db9006", "file": "20_seconds/EgoDex/db1f34a08fbece0550db9006_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "test/assemble_disassemble_structures/0", "source_first_frame": 264, "source_last_frame": 861, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "db1f34a08fbece0550db9006", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EgoDex/db1f34a08fbece0550db9006_P3R3_FIT_20260921_20.00s.mp4", "bytes": 3009900, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "20_seconds/EgoDex/5afc3a04e19ce5ed3615239e_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EgoDex/db1f34a08fbece0550db9006_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "239947f8194093a24d8a5ad2a5ee81387cb7e06c2ad9dc237398cb26f6401e84", "mano_sha256": "3bd8aaafedf83fbb262220ff29d80e0fc8a0a1d564febce58b5f1ebf20b2ae45", "duration_s": 3.0, "frames": 30, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "e922229c382290d3b65c9bd1", "sample_id": "e922229c382290d3b65c9bd1", "file": "short_originals_under_20s/EgoDex/e922229c382290d3b65c9bd1_P3R3_FIT_20260921_3.00s.mp4", "recording_id": "test/pick_place_food/45", "source_first_frame": 0, "source_last_frame": 87, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "no valid P3R2_PREDICTED mesh in 1/30 output frames; may include genuine out-of-view frames", "id": "e922229c382290d3b65c9bd1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/e922229c382290d3b65c9bd1_P3R3_FIT_20260921_3.00s.mp4", "bytes": 595336, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/e922229c382290d3b65c9bd1_P3R2_PREDICTED_3.00s.mp4", "short_originals_under_20s/EgoDex/e922229c382290d3b65c9bd1_P3R3_PREDICTED_3.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "0240b31e74525897be72fbc093662a22ad0e327d242919f925ecb51f44002e4e", "mano_sha256": "1262cb2035c339e9dab627b9b33995ed4643837a019aab2732bbaeeeb37b126f", "duration_s": 7.0, "frames": 70, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "f8466de49aedb6505e5ae1bd", "sample_id": "f8466de49aedb6505e5ae1bd", "file": "short_originals_under_20s/EgoDex/f8466de49aedb6505e5ae1bd_P3R3_FIT_20260921_7.00s.mp4", "recording_id": "test/scoop_dump_ice/2", "source_first_frame": 0, "source_last_frame": 207, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "f8466de49aedb6505e5ae1bd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/EgoDex/f8466de49aedb6505e5ae1bd_P3R3_FIT_20260921_7.00s.mp4", "bytes": 1096625, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "short_originals_under_20s/EgoDex/80d31210ac253d3b0aa0df22_P3R2_PREDICTED_7.00s.mp4", "short_originals_under_20s/EgoDex/f8466de49aedb6505e5ae1bd_P3R3_PREDICTED_7.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EgoDex", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "32f80725afd29dc84d56b821318cab9893f6dc50d168fa1b5fa2470321d2ee4b", "mano_sha256": "fe6d505c3d7a51527397a9944c145011634bc5f8eb3d96cb851fa31099094ab8", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "native source camera trajectory", "intrinsics_source": "source calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "fd9f63972f797dc1ce8a9fdc", "sample_id": "fd9f63972f797dc1ce8a9fdc", "file": "20_seconds/EgoDex/fd9f63972f797dc1ce8a9fdc_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "test/stack_unstack_tupperware/5", "source_first_frame": 1, "source_last_frame": 598, "source_split": "official_test", "training_label_source": "ARKit skeletal estimates, explicit anatomical name mapping; native joints also retained, not MANO", "quality_warnings": "", "id": "fd9f63972f797dc1ce8a9fdc", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EgoDex/fd9f63972f797dc1ce8a9fdc_P3R3_FIT_20260921_20.00s.mp4", "bytes": 3147716, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考为 ARKit 估计；本页暂不纳入可靠参考 motion 筛选，不将它当独立 GT。" }, "camera_display_source": "ARKit device trajectory estimate（不是独立 mocap GT）", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "官方 ARKit 估计", "kind": "estimated", "detail": "数据集发布的 ARKit 设备位姿估计，不是我们运行生成，也不是独立测量 GT。" } }, "superseded_files": [ "20_seconds/EgoDex/020c8da18328b749ccd81351_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EgoDex/fd9f63972f797dc1ce8a9fdc_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "04b959961377b9f9164e1415ab0ea8fa0b9acd5da6a1bba3a6f68d7c848a6855", "mano_sha256": "f5f088e37846f91a987c12ffe09571b9b08e0cdf624502ab49d740c79a44f544", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "047c295e0ac4f2098479eb84", "sample_id": "047c295e0ac4f2098479eb84", "file": "20_seconds/EPIC-KITCHENS/047c295e0ac4f2098479eb84_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P32_07", "source_first_frame": 537, "source_last_frame": 1731, "source_split": "official_validation", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 6/200 output frames; may include genuine out-of-view frames", "id": "047c295e0ac4f2098479eb84", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/047c295e0ac4f2098479eb84_P3R3_FIT_20260921_20.02s.mp4", "bytes": 6257793, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/323101d382e65a7c74878686_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/89fc766f98d5bcf8de0c8504_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/047c295e0ac4f2098479eb84_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "f839b5a0dfc5385197d16d5d0633a669a5fe5864cd855ec99b4674898f37a8f0", "mano_sha256": "f2323e7ea60539be2fbac2662efd9393608090fdfa60db6eaf9fdf2a448e9e20", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "1e147c07e7e038dab54811ac", "sample_id": "1e147c07e7e038dab54811ac", "file": "20_seconds/EPIC-KITCHENS/1e147c07e7e038dab54811ac_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P28_23", "source_first_frame": 2, "source_last_frame": 1196, "source_split": "official_validation", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 23/200 output frames; may include genuine out-of-view frames", "id": "1e147c07e7e038dab54811ac", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/1e147c07e7e038dab54811ac_P3R3_FIT_20260921_20.02s.mp4", "bytes": 5836001, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/f74e7c3b7934f36b2c6e6fbf_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/1e147c07e7e038dab54811ac_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "9a86e34ff2c74573801cf08da34538b050ffde9bb7978ca6b3bc1746b834db6a", "mano_sha256": "0326eabb81020a1ab0adc58e71352cbca72822dc4704486ae258eb3c694f29d1", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "2d54ec5893b955e81670a593", "sample_id": "2d54ec5893b955e81670a593", "file": "20_seconds/EPIC-KITCHENS/2d54ec5893b955e81670a593_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "P02_127", "source_first_frame": 317, "source_last_frame": 1312, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "hand_jump_over_25cm_per_output_frame; no valid P3R2_PREDICTED mesh in 19/200 output frames; may include genuine out-of-view frames", "id": "2d54ec5893b955e81670a593", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/2d54ec5893b955e81670a593_P3R3_FIT_20260921_20.00s.mp4", "bytes": 6400093, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/a9656c7bd824a377a3a47387_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EPIC-KITCHENS/2d54ec5893b955e81670a593_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "79f8220dbedfe77cf2223acfebf216a45d6981583b4a1802b77b8a52519684c7", "mano_sha256": "f7dd28638731b82e2a14173b606b0f85cef7eec900e8a8c2f8a71b6fd389c0ee", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "342a38dd8dd791017ebfd356", "sample_id": "342a38dd8dd791017ebfd356", "file": "20_seconds/EPIC-KITCHENS/342a38dd8dd791017ebfd356_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P11_07", "source_first_frame": 195, "source_last_frame": 1389, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 10/200 output frames; may include genuine out-of-view frames", "id": "342a38dd8dd791017ebfd356", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/342a38dd8dd791017ebfd356_P3R3_FIT_20260921_20.02s.mp4", "bytes": 6063527, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/2c8725b6bec0363616210474_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/342a38dd8dd791017ebfd356_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "5b0699af5a06058644add327f010a74aaedcd6c85fbdf3ac41b4bce9b4c87a84", "mano_sha256": "eda9856f1d6a5b1d231751aadae7145bdf0d555b190bf8a0556de19e41422894", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "408ee054a30602a350c77f03", "sample_id": "408ee054a30602a350c77f03", "file": "20_seconds/EPIC-KITCHENS/408ee054a30602a350c77f03_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P15_11", "source_first_frame": 1, "source_last_frame": 1195, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "408ee054a30602a350c77f03", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/408ee054a30602a350c77f03_P3R3_FIT_20260921_20.02s.mp4", "bytes": 6268877, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/601af4c329629bb7cea4b5fc_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/408ee054a30602a350c77f03_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "c9ebaefc100c421023f762555a0c11a8dfdaa79e11edd21133d9f08157790d4e", "mano_sha256": "40167d0a4119ebfcef709a58e98ad72a61053cf6f664d5916faf2b02c9290def", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "4b6e2e9891cb6056dc5c6ce8", "sample_id": "4b6e2e9891cb6056dc5c6ce8", "file": "20_seconds/EPIC-KITCHENS/4b6e2e9891cb6056dc5c6ce8_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P28_23", "source_first_frame": 154, "source_last_frame": 1348, "source_split": "official_validation", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 32/200 output frames; may include genuine out-of-view frames", "id": "4b6e2e9891cb6056dc5c6ce8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/4b6e2e9891cb6056dc5c6ce8_P3R3_FIT_20260921_20.02s.mp4", "bytes": 6068762, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/ca2f11cd145cca7490450a0f_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/4b6e2e9891cb6056dc5c6ce8_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "e31a4f1add5a729f2d882b7e76ae15cdb79dedfc53d0b8a419b6d88c12e312ed", "mano_sha256": "9cc20617108c6ab517b921ef79bf21ba4aadb240c96f53bf4e016085cf23088b", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "4f31808b6645e08809748b69", "sample_id": "4f31808b6645e08809748b69", "file": "20_seconds/EPIC-KITCHENS/4f31808b6645e08809748b69_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "P03_114", "source_first_frame": 3, "source_last_frame": 998, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 97/200 output frames; may include genuine out-of-view frames", "id": "4f31808b6645e08809748b69", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/4f31808b6645e08809748b69_P3R3_FIT_20260921_20.00s.mp4", "bytes": 5484448, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/16e7c4a14e7f33d6e1025429_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EPIC-KITCHENS/4f31808b6645e08809748b69_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "e491b0aeba7fb507f29f88cd0c513afc0ab7aaf2dee7b792afbaf249e2c28173", "mano_sha256": "382f37a3ec028284bc6d16dd2df7175c3af685eb619c20202b2fa620dd253b33", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "518147b98a7e4913de1ec31a", "sample_id": "518147b98a7e4913de1ec31a", "file": "20_seconds/EPIC-KITCHENS/518147b98a7e4913de1ec31a_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "P26_109", "source_first_frame": 290, "source_last_frame": 1285, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 112/200 output frames; may include genuine out-of-view frames", "id": "518147b98a7e4913de1ec31a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/518147b98a7e4913de1ec31a_P3R3_FIT_20260921_20.00s.mp4", "bytes": 3970827, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/e88a70db79a4af7080e3bde8_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EPIC-KITCHENS/518147b98a7e4913de1ec31a_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "50d0b733a77a594b79dc929b503345f275ce88108460baa305d16919432c09b6", "mano_sha256": "459d6749f0c4cc3c5a36d2806fc035ae4a01e410fe8582b5be219dfc382dfc9d", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "538b11889b3296fe152d49d4", "sample_id": "538b11889b3296fe152d49d4", "file": "20_seconds/EPIC-KITCHENS/538b11889b3296fe152d49d4_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P15_11", "source_first_frame": 447, "source_last_frame": 1641, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "hand_jump_over_25cm_per_output_frame", "id": "538b11889b3296fe152d49d4", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/538b11889b3296fe152d49d4_P3R3_FIT_20260921_20.02s.mp4", "bytes": 5788685, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/d0140245dc6dbcfd76804378_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/538b11889b3296fe152d49d4_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "b92a2309cd720dd3c9eb541337d92f804bcf8c488d3ef7aef7c2676056f7ba9f", "mano_sha256": "18b5ce75fc8d72162b921687ddfc5b86c59e26423e0f9c857fb7618b24226a47", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "614c5b6ed0772c7a13b8c960", "sample_id": "614c5b6ed0772c7a13b8c960", "file": "20_seconds/EPIC-KITCHENS/614c5b6ed0772c7a13b8c960_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P32_07", "source_first_frame": 5, "source_last_frame": 1199, "source_split": "official_validation", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 5/200 output frames; may include genuine out-of-view frames", "id": "614c5b6ed0772c7a13b8c960", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/614c5b6ed0772c7a13b8c960_P3R3_FIT_20260921_20.02s.mp4", "bytes": 6457192, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/cb0ff184c2d340a22ca03084_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/614c5b6ed0772c7a13b8c960_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "543a4253e2d17c78e2c00409f58ae7804bff176bd9d8a869a8353065dbfe2d1d", "mano_sha256": "bf64290573f282f54a769aa390bf9c2a435e0d81c862814143426840f483efa4", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "647d62051f6353949bb6e5cd", "sample_id": "647d62051f6353949bb6e5cd", "file": "20_seconds/EPIC-KITCHENS/647d62051f6353949bb6e5cd_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P15_11", "source_first_frame": 224, "source_last_frame": 1418, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "hand_jump_over_25cm_per_output_frame", "id": "647d62051f6353949bb6e5cd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/647d62051f6353949bb6e5cd_P3R3_FIT_20260921_20.02s.mp4", "bytes": 5552956, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/2dabf15fa741d1d806fa6faf_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/647d62051f6353949bb6e5cd_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "e6776ce8758f834863954491adf8eca08c7ebdb2151cfbc1d0df4adfab912949", "mano_sha256": "993348ac91b92978b36318e58a78342f7e1b302b693bae8c042082c2279dbe43", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "683c16466ea81a2a513aa4e7", "sample_id": "683c16466ea81a2a513aa4e7", "file": "20_seconds/EPIC-KITCHENS/683c16466ea81a2a513aa4e7_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P28_23", "source_first_frame": 306, "source_last_frame": 1500, "source_split": "official_validation", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 45/200 output frames; may include genuine out-of-view frames", "id": "683c16466ea81a2a513aa4e7", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/683c16466ea81a2a513aa4e7_P3R3_FIT_20260921_20.02s.mp4", "bytes": 6376355, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/d937b2d9e565ad40de9c24fe_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/683c16466ea81a2a513aa4e7_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "3ff70278ad7a2aa45660e6d98591346c022a74748c0ccd191cc6345dd73867f5", "mano_sha256": "f2737a49a031cc060492579df5b84df09f63d109e9b525fd27818d9b63a263b9", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "6f1179fdbd9d66506ae2fbff", "sample_id": "6f1179fdbd9d66506ae2fbff", "file": "20_seconds/EPIC-KITCHENS/6f1179fdbd9d66506ae2fbff_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P14_02", "source_first_frame": 367, "source_last_frame": 1561, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 12/200 output frames; may include genuine out-of-view frames", "id": "6f1179fdbd9d66506ae2fbff", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/6f1179fdbd9d66506ae2fbff_P3R3_FIT_20260921_20.02s.mp4", "bytes": 4949789, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/b8b68a4b5ec3bba29c5c16c5_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/6f1179fdbd9d66506ae2fbff_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "aa4fa51f2a26063d683b61ad711bb1fef774c72994380e0a4dead207d5399bd7", "mano_sha256": "a9a618d8167967ddadde2d204fbd64adb7427ddfd90e17812a26c6e1507c85cc", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "84193664737944669f3eb95c", "sample_id": "84193664737944669f3eb95c", "file": "20_seconds/EPIC-KITCHENS/84193664737944669f3eb95c_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "P03_114", "source_first_frame": 0, "source_last_frame": 995, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "hand_jump_over_25cm_per_output_frame; no valid P3R2_PREDICTED mesh in 98/200 output frames; may include genuine out-of-view frames", "id": "84193664737944669f3eb95c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/84193664737944669f3eb95c_P3R3_FIT_20260921_20.00s.mp4", "bytes": 5514857, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/e35ced80473e411e6d53b823_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EPIC-KITCHENS/84193664737944669f3eb95c_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "5cecad6b06f66c301cb88f3babc6920a6ab70acda5e8235fc144a6f20583d48f", "mano_sha256": "02003e9f60eac25a51d386f18eb1a5443405ff856d3be51559db8f9c63dc312b", "duration_s": 20.0, "frames": 200, "fps": 10.0, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "95a33b191b103abbd8ff1583", "sample_id": "95a33b191b103abbd8ff1583", "file": "20_seconds/EPIC-KITCHENS/95a33b191b103abbd8ff1583_P3R3_FIT_20260921_20.00s.mp4", "recording_id": "P26_109", "source_first_frame": 0, "source_last_frame": 995, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 134/200 output frames; may include genuine out-of-view frames", "id": "95a33b191b103abbd8ff1583", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/95a33b191b103abbd8ff1583_P3R3_FIT_20260921_20.00s.mp4", "bytes": 4362854, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 10.000 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/1d326f3058bda1da335b4d3b_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EPIC-KITCHENS/a072e7d7a187d36f154ff745_P3R2_PREDICTED_20.00s.mp4", "20_seconds/EPIC-KITCHENS/95a33b191b103abbd8ff1583_P3R3_PREDICTED_20.00s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "a58dabf6fbae8da919a071a44a941e394fe479f72505a4bbb4a596cbfd50c32d", "mano_sha256": "1b1105d0012395bd1bcd3b13ffec1a60a42a33f8619166c70c5ce53fe015b0aa", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "a4af1873f795eec4dcbd4a69", "sample_id": "a4af1873f795eec4dcbd4a69", "file": "20_seconds/EPIC-KITCHENS/a4af1873f795eec4dcbd4a69_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P14_02", "source_first_frame": 0, "source_last_frame": 1194, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 24/200 output frames; may include genuine out-of-view frames", "id": "a4af1873f795eec4dcbd4a69", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/a4af1873f795eec4dcbd4a69_P3R3_FIT_20260921_20.02s.mp4", "bytes": 4814981, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/85bdcbf36ff1a8da52d295c1_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/cebb175135eb8d6deb59244e_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/a4af1873f795eec4dcbd4a69_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "13e141d8e11c06a9a7adb21caf326a672d222509a160a20b4b2785cfb4879c60", "mano_sha256": "bf2e6b67d424d46a2ce6a0caffceafd89c653daee983bfb4652cd9cc644f474a", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "a5974b1bcac1d3a78633c7d7", "sample_id": "a5974b1bcac1d3a78633c7d7", "file": "20_seconds/EPIC-KITCHENS/a5974b1bcac1d3a78633c7d7_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P32_07", "source_first_frame": 0, "source_last_frame": 1194, "source_split": "official_validation", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 14/200 output frames; may include genuine out-of-view frames", "id": "a5974b1bcac1d3a78633c7d7", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/a5974b1bcac1d3a78633c7d7_P3R3_FIT_20260921_20.02s.mp4", "bytes": 6361109, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/83be9cc8b73b12370c0feaa4_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/a5974b1bcac1d3a78633c7d7_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "5c717f6513d5678c399fc58a1ca64dc1cec289f337f4289f21820227e1a57bcc", "mano_sha256": "98266c503d96cd1cb5977c88da41784005a26f7a6f3a25b6b83431e5aba87b60", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "a6c5b7f072b31801c8841a92", "sample_id": "a6c5b7f072b31801c8841a92", "file": "20_seconds/EPIC-KITCHENS/a6c5b7f072b31801c8841a92_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P14_02", "source_first_frame": 735, "source_last_frame": 1929, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 53/200 output frames; may include genuine out-of-view frames", "id": "a6c5b7f072b31801c8841a92", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/a6c5b7f072b31801c8841a92_P3R3_FIT_20260921_20.02s.mp4", "bytes": 4711476, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/9f56f7bd7701bd46dd7aa042_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/a6c5b7f072b31801c8841a92_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "67ce5e118bac869ce4445982a9364f9b3d7060deb511608d84da8b8360a4432e", "mano_sha256": "4b26c7af8a8e7f413e9876a92f9f570059ec5e2e33cc1f2096c85425709378c8", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "bb9ab5bf73eea598ee20572d", "sample_id": "bb9ab5bf73eea598ee20572d", "file": "20_seconds/EPIC-KITCHENS/bb9ab5bf73eea598ee20572d_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P14_02", "source_first_frame": 734, "source_last_frame": 1928, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 51/200 output frames; may include genuine out-of-view frames", "id": "bb9ab5bf73eea598ee20572d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/bb9ab5bf73eea598ee20572d_P3R3_FIT_20260921_20.02s.mp4", "bytes": 4590865, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/3a0c8d088df915bc3037798e_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/bb9ab5bf73eea598ee20572d_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "188de01276964ffaee49327a41925a33b99d8b46a4b9cc63d7d07f8677e09a7d", "mano_sha256": "632dcadfbf61ca30623028de357fc5e86d252d70204b2fb4b6c825475a1ac176", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "cb06359e66d342bdf285007e", "sample_id": "cb06359e66d342bdf285007e", "file": "20_seconds/EPIC-KITCHENS/cb06359e66d342bdf285007e_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P06_08", "source_first_frame": 0, "source_last_frame": 1194, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "hand_jump_over_25cm_per_output_frame; no valid P3R2_PREDICTED mesh in 102/200 output frames; may include genuine out-of-view frames", "id": "cb06359e66d342bdf285007e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/cb06359e66d342bdf285007e_P3R3_FIT_20260921_20.02s.mp4", "bytes": 5912630, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/c39f43c79bef49f3d9ae8ad0_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/cb06359e66d342bdf285007e_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "9c423aae2b16d9e4ef2838c4e0c15cbe2c43e07b25d2d0719ab02cbca3c018ef", "mano_sha256": "e1eded29fb0b140c83484de63d576116191756e3f1488ede09cd7fc928c6918a", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "d10ba85e3a6db264be66aa15", "sample_id": "d10ba85e3a6db264be66aa15", "file": "20_seconds/EPIC-KITCHENS/d10ba85e3a6db264be66aa15_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P08_12", "source_first_frame": 823, "source_last_frame": 2017, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 44/200 output frames; may include genuine out-of-view frames", "id": "d10ba85e3a6db264be66aa15", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/d10ba85e3a6db264be66aa15_P3R3_FIT_20260921_20.02s.mp4", "bytes": 4723869, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/6cde562ac5098bc7120f114b_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/d10ba85e3a6db264be66aa15_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "c21cad3a57ac9a4c7262a7275cdc83ca78c32fb55fe8f56ebeea854c1b5195c5", "mano_sha256": "6d4f666eb1b5c2be84ba1cd85e6bb8633280826d94463a17b28e8a18e1220797", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "d529c02179db68b72fe72a73", "sample_id": "d529c02179db68b72fe72a73", "file": "20_seconds/EPIC-KITCHENS/d529c02179db68b72fe72a73_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P06_08", "source_first_frame": 638, "source_last_frame": 1832, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 90/200 output frames; may include genuine out-of-view frames", "id": "d529c02179db68b72fe72a73", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/d529c02179db68b72fe72a73_P3R3_FIT_20260921_20.02s.mp4", "bytes": 5796752, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/d2c11c690dc18a9383fbabf6_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/d529c02179db68b72fe72a73_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "6a95b48ab2c0788b9d97c402f2ec241c918b2461205a02b3dc009c25c5551f52", "mano_sha256": "d7f57e8e6cd1f90195ac1efcf29a20844a115cca5371822eb5c1234f19a89060", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "ebb8cbece6b663b95a81987b", "sample_id": "ebb8cbece6b663b95a81987b", "file": "20_seconds/EPIC-KITCHENS/ebb8cbece6b663b95a81987b_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P08_12", "source_first_frame": 822, "source_last_frame": 2016, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 40/200 output frames; may include genuine out-of-view frames", "id": "ebb8cbece6b663b95a81987b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/ebb8cbece6b663b95a81987b_P3R3_FIT_20260921_20.02s.mp4", "bytes": 4812069, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/0157f95d69bb78a574e69e98_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/ebb8cbece6b663b95a81987b_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "844a9b83767ebc001ef0536a1ea5d08d9fa677fe3be94634405f65eeadb67dec", "mano_sha256": "2dc733e81cea7776fb9c24f17852b2ba4e199334345e0a06e4eeb4af480b01bd", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "ef9998c43e769d9c351276f4", "sample_id": "ef9998c43e769d9c351276f4", "file": "20_seconds/EPIC-KITCHENS/ef9998c43e769d9c351276f4_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P11_07", "source_first_frame": 4, "source_last_frame": 1198, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 25/200 output frames; may include genuine out-of-view frames", "id": "ef9998c43e769d9c351276f4", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/ef9998c43e769d9c351276f4_P3R3_FIT_20260921_20.02s.mp4", "bytes": 5224195, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/c3486dab73e7181093947ae2_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/ef9998c43e769d9c351276f4_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "5a3d5ce1f88fcfde37cdb6caeea29a73501f8e05b4a2844eca2cad37dfc94b98", "mano_sha256": "74f05d71de1d88035658a11275580e07a1241fa5edbd904a002d69da1314398b", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "f0591f9fd1551f63b4390c40", "sample_id": "f0591f9fd1551f63b4390c40", "file": "20_seconds/EPIC-KITCHENS/f0591f9fd1551f63b4390c40_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P15_11", "source_first_frame": 449, "source_last_frame": 1643, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "f0591f9fd1551f63b4390c40", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/f0591f9fd1551f63b4390c40_P3R3_FIT_20260921_20.02s.mp4", "bytes": 5810981, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/53a4460bfa9f94bd8d95995f_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/f0591f9fd1551f63b4390c40_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "fc1af0def4e3d2029d90e18ec047f95a53e791d2b1ba926516d7ac940decf6ce", "mano_sha256": "421d7917394db861499d1ca10f5249ba5f050764b612995e2f767ffbab9045d6", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "f0ee742808bfce93b2e52d41", "sample_id": "f0ee742808bfce93b2e52d41", "file": "20_seconds/EPIC-KITCHENS/f0ee742808bfce93b2e52d41_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P15_11", "source_first_frame": 0, "source_last_frame": 1194, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "f0ee742808bfce93b2e52d41", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/f0ee742808bfce93b2e52d41_P3R3_FIT_20260921_20.02s.mp4", "bytes": 6096255, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/1024c6b99d084c170449a5fb_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/f0ee742808bfce93b2e52d41_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "EPIC-KITCHENS", "source_tag": "P3R3_PREDICTED", "variant": "v9p3r3", "video_sha256": "5aa46dbffb98c70810e0019ab555780c50cfa57842450d380c719a892f3fff2d", "mano_sha256": "9ec6e1b9e9c467b33e4030d7129326c0e45b23963c65e41cc938f8b6cbacc67f", "duration_s": 20.02, "frames": 200, "fps": 9.99000999000999, "mano_source": "P3R3 管线自造 MANO；恢复受约束的手形/姿态/深度/轮廓拟合，surface_fit_restore_20260921；未使用手部 GT", "scene_depth_source": "DA3METRIC-LARGE predicted RGB depth; learned metric scale, not sensor GT", "scene_camera_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "intrinsics_source": "GeoCalib predicted intrinsics, not calibration", "scene_scale_anchor": "one sequence-wide robust MANO-to-DA3 gauge; V9-P3R2 predicted MANO; visual scale anchor, not GT", "manual_note": "本片段已重新拟合与渲染；旧版人工审核不自动继承。仅作预测可视化，不是训练 GT。", "previous_visual_review_hash_matched": false, "technical_pass": true, "training_label_decision": "not_GT_prediction_only", "usage": "visualization_only", "context_id": "f135afcd8e062400b420cf33", "sample_id": "f135afcd8e062400b420cf33", "file": "20_seconds/EPIC-KITCHENS/f135afcd8e062400b420cf33_P3R3_FIT_20260921_20.02s.mp4", "recording_id": "P08_12", "source_first_frame": 411, "source_last_frame": 1605, "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "no valid P3R2_PREDICTED mesh in 8/200 output frames; may include genuine out-of-view frames", "id": "f135afcd8e062400b420cf33", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/EPIC-KITCHENS/f135afcd8e062400b420cf33_P3R3_FIT_20260921_20.02s.mp4", "bytes": 5176856, "width": 1710, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "未提供手部与相机共同世界系参考；视频使用预测 MANO 和估计轨迹。" }, "camera_display_source": "historical ViPE using predicted DA3 depth; 9.990 Hz poses, SLERP to RGB timestamps with endpoint hold", "mano_display_kind": "管线自造", "sequence_mapping": "verified original source frame map; exact recording ID", "provenance": { "mano": { "label": "管线自造", "kind": "predicted", "detail": "修复后的 P3R3 管线自造；恢复受约束的手形/姿态/深度和轮廓拟合，仍不是 MANO GT。" }, "pointcloud": { "label": "管线自造", "kind": "predicted", "detail": "由 DA3 预测深度并融合生成，不是传感器实测点云。" }, "camera": { "label": "管线自造", "kind": "predicted", "detail": "由我们的 ViPE 路线估计相机轨迹，不是官方相机 GT。" } }, "superseded_files": [ "20_seconds/EPIC-KITCHENS/a4a9ff925b7ebbfb227e31da_P3R2_PREDICTED_20.02s.mp4", "20_seconds/EPIC-KITCHENS/f135afcd8e062400b420cf33_P3R3_PREDICTED_20.02s.mp4" ], "training_allowed": false, "pipeline_revision": "surface_fit_restore_20260921", "visual_review_scope": "本次逐片技术核验；人工抽帧审核范围见交付报告。" }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "439c2fb572296692734386446db89204d0c999e478d26cbb804f83a43c3866c6", "mano_sha256": "5b3069c1a096010e24d79c22e49d9889aedd2597e934ad2eb54feaca3f361cce", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "11f91126fa8103c575e1f813", "sample_id": "11f91126fa8103c575e1f813", "file": "20_seconds/HO3D/11f91126fa8103c575e1f813_MANO_GT_20.00s.mp4", "recording_id": "SiBF14", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "11f91126fa8103c575e1f813", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/11f91126fa8103c575e1f813_MANO_GT_20.00s.mp4", "bytes": 5064868, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/11f91126fa8103c575e1f813/reference.json", "sha256": "6e80718004f1b04550aa5de29037163122d5b71d58b0f686c47f90ddc5b5b626" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d342f21bfce3920db3efc58763ab88b9575e522200ffcbf180bd4eb4239c2f19", "mano_sha256": "9c4a27fa39811d76f470753cb06212e0b9606e2b02d408758bcd8b22eccd25c3", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "1839da6154b96a159a272d0c", "sample_id": "1839da6154b96a159a272d0c", "file": "20_seconds/HO3D/1839da6154b96a159a272d0c_MANO_GT_20.00s.mp4", "recording_id": "SMu41", "source_first_frame": "600", "source_last_frame": "1199", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "1839da6154b96a159a272d0c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/1839da6154b96a159a272d0c_MANO_GT_20.00s.mp4", "bytes": 3944950, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/1839da6154b96a159a272d0c/reference.json", "sha256": "373f8126bf65ff147bf6f685af8807d7750f86bc9e0f795f92ae6eb3ad4cc38f" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d9bc837074aca7f83e06bffe8eab90b94997585e52c28fdb2472699c214c6b4e", "mano_sha256": "60ea3ec88c07dd1d615f45ab574a61a45de7a71af23adf5aeeecbff05a8e4238", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "官方拟合 MANO 保留腕部截面与手物接触偏差，未用模型替换。", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "1f0648eacd49719be89fe80b", "sample_id": "1f0648eacd49719be89fe80b", "file": "20_seconds/HO3D/1f0648eacd49719be89fe80b_MANO_GT_20.00s.mp4", "recording_id": "BB10", "source_first_frame": "413", "source_last_frame": "1012", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "1f0648eacd49719be89fe80b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/1f0648eacd49719be89fe80b_MANO_GT_20.00s.mp4", "bytes": 4497001, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/1f0648eacd49719be89fe80b/reference.json", "sha256": "66fe0ab153c78f452e6d22ed13daa082a8139e13eed54f6ccc702a45a28d9e12" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "de0346d59c74c57dd9d7d8198d9ac727d2a87a4ddd2c32fc30c9d7670a09fb06", "mano_sha256": "855c043e28eeb4bfef250fd07e736b7466441465f060653d2f71048b2ceecaa3", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "24b100fa3b506fac5ea48c94", "sample_id": "24b100fa3b506fac5ea48c94", "file": "20_seconds/HO3D/24b100fa3b506fac5ea48c94_MANO_GT_20.00s.mp4", "recording_id": "SiBF12", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "24b100fa3b506fac5ea48c94", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/24b100fa3b506fac5ea48c94_MANO_GT_20.00s.mp4", "bytes": 4603673, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/24b100fa3b506fac5ea48c94/reference.json", "sha256": "b083475537188c058608ca478894cbac69caec78823458d399c7d1ea9319b0ce" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8deb73c5f787a4aeb11535b183bfa2bac26db9a58a6013c8b7c7fd24132d5149", "mano_sha256": "a6818d89852c0a171f8d1df438a3cfaaf1ae46359c5b423f5251001059bd8faf", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "26471dcb6c428f3d279b30ee", "sample_id": "26471dcb6c428f3d279b30ee", "file": "20_seconds/HO3D/26471dcb6c428f3d279b30ee_MANO_GT_20.00s.mp4", "recording_id": "SS3", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "26471dcb6c428f3d279b30ee", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/26471dcb6c428f3d279b30ee_MANO_GT_20.00s.mp4", "bytes": 5523246, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/26471dcb6c428f3d279b30ee/reference.json", "sha256": "b1374d6062e9df3c4a79b806f9792bc51502e917d385a4f6af97305230c13086" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6c8a9be04025a67ebb542672533d084dcd10ee2b90ec169f2d455caf65e30648", "mano_sha256": "51976cb0f0357494c22b7965e994e9fc88a40c2262abecb473202cb8e53af64f", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "29fde70eefb3dd4350968483", "sample_id": "29fde70eefb3dd4350968483", "file": "20_seconds/HO3D/29fde70eefb3dd4350968483_MANO_GT_20.00s.mp4", "recording_id": "SiBF10", "source_first_frame": "1479", "source_last_frame": "2078", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "29fde70eefb3dd4350968483", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/29fde70eefb3dd4350968483_MANO_GT_20.00s.mp4", "bytes": 4754192, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/29fde70eefb3dd4350968483/reference.json", "sha256": "dd4e00aa28ecd00e1de24fe28563f20a8d0ddfafb69c6caca5a9c1508cc1ca5d" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "98c7f72e69d4f29372722b3438962f2ee622db540d6cccf27cb63ea6c2111578", "mano_sha256": "e726305081267c0af1e324546ecbbda7d2b0c889e26907bad651f9e2f3ecfdfb", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "2a0d36dbd9f5ed3c513ef8c2", "sample_id": "2a0d36dbd9f5ed3c513ef8c2", "file": "20_seconds/HO3D/2a0d36dbd9f5ed3c513ef8c2_MANO_GT_20.00s.mp4", "recording_id": "SB12", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "2a0d36dbd9f5ed3c513ef8c2", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/2a0d36dbd9f5ed3c513ef8c2_MANO_GT_20.00s.mp4", "bytes": 5055845, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/2a0d36dbd9f5ed3c513ef8c2/reference.json", "sha256": "2638d175af8fd3fd74653c9d325d14fe125f607d29caf65a211c3cbf55cbeb4a" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5cb610581c14740705798e34bcc48013ec293a1a9eb5da06f001c0fc04015adc", "mano_sha256": "da675c60f5461d049a446400efffb53a0535fb9bccabd89f56aba7a94d32ee72", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "2c37844b5ac5b6f0fee08ba3", "sample_id": "2c37844b5ac5b6f0fee08ba3", "file": "20_seconds/HO3D/2c37844b5ac5b6f0fee08ba3_MANO_GT_20.00s.mp4", "recording_id": "ShSu14", "source_first_frame": "21", "source_last_frame": "620", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "2c37844b5ac5b6f0fee08ba3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/2c37844b5ac5b6f0fee08ba3_MANO_GT_20.00s.mp4", "bytes": 4881111, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/2c37844b5ac5b6f0fee08ba3/reference.json", "sha256": "ba8415aac072891d9e1653272b80f21938c779f1efe57b3b5c7bcceb6b9cece8" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "9c9a70dea0146e6740f66b92139c6438cd7671a185bd160319c2b0992e4377c4", "mano_sha256": "df5ee54295333127e90e43380efb1522e5dc5dd3af2ea824b2bced6a3dac8184", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "2d7a1870d53400e69466c8d6", "sample_id": "2d7a1870d53400e69466c8d6", "file": "20_seconds/HO3D/2d7a1870d53400e69466c8d6_MANO_GT_20.00s.mp4", "recording_id": "SMu42", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "2d7a1870d53400e69466c8d6", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/2d7a1870d53400e69466c8d6_MANO_GT_20.00s.mp4", "bytes": 4332225, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/2d7a1870d53400e69466c8d6/reference.json", "sha256": "6ca73d00faaee3c11e1f38dcc92f3685ee73d0348dfb19543e54c9b28fd5bcdc" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "738ed7ead8512e06b88920dd69eaab3ed0868be58ff2c5ee7e24cf455df0ea98", "mano_sha256": "767b60539c268511495c38ad0ac3b5068af123b659704fc5f4dceccb22120af0", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "360e337159a4d21bc4ca73a7", "sample_id": "360e337159a4d21bc4ca73a7", "file": "20_seconds/HO3D/360e337159a4d21bc4ca73a7_MANO_GT_20.00s.mp4", "recording_id": "GSF11", "source_first_frame": "600", "source_last_frame": "1199", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "360e337159a4d21bc4ca73a7", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/360e337159a4d21bc4ca73a7_MANO_GT_20.00s.mp4", "bytes": 4638653, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/360e337159a4d21bc4ca73a7/reference.json", "sha256": "b9c60e0a99ac5d42ed7ebfc5f443be8641c442aa2bafaaf6ae2cdd8916a59a21" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0f37c6f17abfc1143cbc4395b0a910d63f47c529042ba3d67ce41460208c4c73", "mano_sha256": "d98eb8e3bd24ea21c487308c95fc1e848df796cb6b36fdb4ecec461ae5872990", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "3a050d66b7b5bc3967523abd", "sample_id": "3a050d66b7b5bc3967523abd", "file": "20_seconds/HO3D/3a050d66b7b5bc3967523abd_MANO_GT_20.00s.mp4", "recording_id": "SS2", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "3a050d66b7b5bc3967523abd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/3a050d66b7b5bc3967523abd_MANO_GT_20.00s.mp4", "bytes": 6017264, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/3a050d66b7b5bc3967523abd/reference.json", "sha256": "1494dffb8478b8a50b8ef6a6722902ea6fe015ae9fbd18930c055164416164d0" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6a922ae2f7361d6bc1c3ad3934e422d667c9645961b1f896e42f83fa78f16343", "mano_sha256": "3ced0b8d443811aeec6cef40e42a740644877817deed90f890f99de0f47f1c7a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "3a2c870bb449726b684be5e7", "sample_id": "3a2c870bb449726b684be5e7", "file": "20_seconds/HO3D/3a2c870bb449726b684be5e7_MANO_GT_20.00s.mp4", "recording_id": "ShSu12", "source_first_frame": "21", "source_last_frame": "620", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "3a2c870bb449726b684be5e7", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/3a2c870bb449726b684be5e7_MANO_GT_20.00s.mp4", "bytes": 4898597, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/3a2c870bb449726b684be5e7/reference.json", "sha256": "b386856741369a93fb56c0ab04af2923fa4bde5d93f73fa928cb1d5774c21e4e" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b701ab643e3ca445d3ea87227570b7471b3222e54e6db4910db8ac58f1816879", "mano_sha256": "f781ac39285114e4dd750d0d3f8f9c0703b069622c873f6bce9a16e97e4e160a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "3e1be64ecd6517907388f113", "sample_id": "3e1be64ecd6517907388f113", "file": "20_seconds/HO3D/3e1be64ecd6517907388f113_MANO_GT_20.00s.mp4", "recording_id": "ShSu10", "source_first_frame": "21", "source_last_frame": "620", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "3e1be64ecd6517907388f113", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/3e1be64ecd6517907388f113_MANO_GT_20.00s.mp4", "bytes": 4998915, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/3e1be64ecd6517907388f113/reference.json", "sha256": "ce73fef9d044a819e5c0422ef58d0cb5087c89d452cf6cf2308e3b7f6c4ed5bc" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "9255ced5b73fefcfb42f09d79f2079b7cc660ca5250d48174c19e312ec82f7ba", "mano_sha256": "ec358037a97857c1f8e08969114ed486e1106c571cd38ede3e7a12625ffdda9a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "43929244f4e2cf3bb506b06b", "sample_id": "43929244f4e2cf3bb506b06b", "file": "20_seconds/HO3D/43929244f4e2cf3bb506b06b_MANO_GT_20.00s.mp4", "recording_id": "SB14", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "43929244f4e2cf3bb506b06b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/43929244f4e2cf3bb506b06b_MANO_GT_20.00s.mp4", "bytes": 4930399, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/43929244f4e2cf3bb506b06b/reference.json", "sha256": "c70e11d506531e02b5398a78df5d4903fa73a7aa96283eed7269c560472e19aa" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4e5e1264a381ab6fac2da99fdf695e93f314dc7b69078d27ba227116094f8665", "mano_sha256": "619a6d7304cfee001739758743bcf22a703a42f2c1b4aea531e81763532c7283", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "4c9ddde1ae8218c48a02cdd1", "sample_id": "4c9ddde1ae8218c48a02cdd1", "file": "20_seconds/HO3D/4c9ddde1ae8218c48a02cdd1_MANO_GT_20.00s.mp4", "recording_id": "SS1", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "4c9ddde1ae8218c48a02cdd1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/4c9ddde1ae8218c48a02cdd1_MANO_GT_20.00s.mp4", "bytes": 6059260, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/4c9ddde1ae8218c48a02cdd1/reference.json", "sha256": "4d051187514c0a6c26aeab92552d57478d4a4064acfbd0eb1d9e76dddf655547" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7891859a19d43e2c35f222f10b03d7629535cea9d7a06135538d56821b9cca7f", "mano_sha256": "d107f76aba816b918aa47e3ac9875e1c3f61317cdc2183e50fe9b2821b28ba1d", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "4cacb691509e4ed3d8745e29", "sample_id": "4cacb691509e4ed3d8745e29", "file": "20_seconds/HO3D/4cacb691509e4ed3d8745e29_MANO_GT_20.00s.mp4", "recording_id": "GSF12", "source_first_frame": "600", "source_last_frame": "1199", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "4cacb691509e4ed3d8745e29", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/4cacb691509e4ed3d8745e29_MANO_GT_20.00s.mp4", "bytes": 5569118, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/4cacb691509e4ed3d8745e29/reference.json", "sha256": "65c876c197e0deea12c2d56c65d78019c0ed6791fa9f47855248596dd6eac003" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "14ff9f97c879f61611bcb2dfdf4a1e46ccc0d61ff4ad98d142967bb721f4493f", "mano_sha256": "461017e3043702bac842c8ad5cf9d1048e72fbc124989d0afd47d3a09213cbf5", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "4d89b7450807fae9b6b8a3ea", "sample_id": "4d89b7450807fae9b6b8a3ea", "file": "20_seconds/HO3D/4d89b7450807fae9b6b8a3ea_MANO_GT_20.00s.mp4", "recording_id": "SiBF13", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "4d89b7450807fae9b6b8a3ea", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/4d89b7450807fae9b6b8a3ea_MANO_GT_20.00s.mp4", "bytes": 5025664, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/4d89b7450807fae9b6b8a3ea/reference.json", "sha256": "a164d89a95af070fa0440914c043024e80b74a45ee91e20dc64692c07ac74e70" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e6c43b600e55095de95b1d56a847be6096d19206968ea2d5d2cf5cdfd99e2b89", "mano_sha256": "001c113510cc4b8617c7778d40381c78ea3ef4886642d614467a84fa79fcea16", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "52fb728070ddbdc88a0d0663", "sample_id": "52fb728070ddbdc88a0d0663", "file": "20_seconds/HO3D/52fb728070ddbdc88a0d0663_MANO_GT_20.00s.mp4", "recording_id": "SMu40", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "52fb728070ddbdc88a0d0663", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/52fb728070ddbdc88a0d0663_MANO_GT_20.00s.mp4", "bytes": 4784968, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/52fb728070ddbdc88a0d0663/reference.json", "sha256": "2028029f0850607836bddeca3e7eb5075df74a9f2056036971e3ffa7e45d831b" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "46165ddbf4f76fab0ab8c7a47a043d00bdd7eff5503786b4c44ab70c75d51df5", "mano_sha256": "1cfb2295fceb45c1c79a7b9ae5be6d8e0ebb40b0385eb1a13429d18439759673", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "56b3f064c01d11b97e06125f", "sample_id": "56b3f064c01d11b97e06125f", "file": "20_seconds/HO3D/56b3f064c01d11b97e06125f_MANO_GT_20.00s.mp4", "recording_id": "MC4", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "56b3f064c01d11b97e06125f", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/56b3f064c01d11b97e06125f_MANO_GT_20.00s.mp4", "bytes": 6514289, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/56b3f064c01d11b97e06125f/reference.json", "sha256": "790a97d367e5f40a44edb11ab3669fd6fb080108dd79be53ac298fd223c5fd6d" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ce1874481ca1b554f297913a6f705ea79a758640016291c642042a910f8d9486", "mano_sha256": "7bc8e444bdbd557ebe4909d2be13c58d3da5eb3273e501ae4812a8d2e646d539", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "57280f7b5ce0dd870af889f3", "sample_id": "57280f7b5ce0dd870af889f3", "file": "20_seconds/HO3D/57280f7b5ce0dd870af889f3_MANO_GT_20.00s.mp4", "recording_id": "SMu1", "source_first_frame": "1145", "source_last_frame": "1744", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "57280f7b5ce0dd870af889f3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/57280f7b5ce0dd870af889f3_MANO_GT_20.00s.mp4", "bytes": 3554790, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/57280f7b5ce0dd870af889f3/reference.json", "sha256": "8daa64b56f800f7c04d613256ab2d05260757cf7da152006969cdfd6057a68ce" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "393277c473a9e54a6b2e0186ee0c45c7fdee132a37dd3fdd0a339c3a49503aa2", "mano_sha256": "226fe38a41b37f9de15608fda5e81387edb439e271a7e69550776a81ade0f6a7", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "5acd8dc0cd3ab96e24ae50ae", "sample_id": "5acd8dc0cd3ab96e24ae50ae", "file": "20_seconds/HO3D/5acd8dc0cd3ab96e24ae50ae_MANO_GT_20.00s.mp4", "recording_id": "SiBF13", "source_first_frame": "1479", "source_last_frame": "2078", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "5acd8dc0cd3ab96e24ae50ae", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/5acd8dc0cd3ab96e24ae50ae_MANO_GT_20.00s.mp4", "bytes": 4425296, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/5acd8dc0cd3ab96e24ae50ae/reference.json", "sha256": "0080fda4e80f909b99563856ed83477295e18b55b2abe43532983ec10a9fb36a" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "160528f7ad9726dad80257f23615edc705af05dcdb4247102c3c053bc82868be", "mano_sha256": "cbad1b385e05cc3503667cb706bdb7dd278f0d488c0e597384e689da4e7e23ac", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "5b794891cffb2f0d02b37c0e", "sample_id": "5b794891cffb2f0d02b37c0e", "file": "20_seconds/HO3D/5b794891cffb2f0d02b37c0e_MANO_GT_20.00s.mp4", "recording_id": "GSF13", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "5b794891cffb2f0d02b37c0e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/5b794891cffb2f0d02b37c0e_MANO_GT_20.00s.mp4", "bytes": 7061488, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/5b794891cffb2f0d02b37c0e/reference.json", "sha256": "6303360a03f4b7d2d78ebec7188bf01203c2799a00f8d6d559982d994fd4875b" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a81e15bbb70358208572558156b1f08d2fcbad4b17678aa4d9d1c5b9597b25d4", "mano_sha256": "a9b552528dc6c56ecd2fbcf13799bc11ad8cb90fc40a78332100010cdaad5fdf", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "5c5baa73e4b5e2739614e6e0", "sample_id": "5c5baa73e4b5e2739614e6e0", "file": "20_seconds/HO3D/5c5baa73e4b5e2739614e6e0_MANO_GT_20.00s.mp4", "recording_id": "GSF14", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "5c5baa73e4b5e2739614e6e0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/5c5baa73e4b5e2739614e6e0_MANO_GT_20.00s.mp4", "bytes": 4681935, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/5c5baa73e4b5e2739614e6e0/reference.json", "sha256": "809ebcf2c80d4f1b50add175339d1c37f218bbd4a6381cb0426522ff99e06fcc" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "db9194d8afa1e3b4c567159aff8347695820b8ae1b0de7f7690600feae5ea0f5", "mano_sha256": "d581554a072a14d9e55a34ed2fd43f658d22a737e7409a18fc6b8a08b947fad3", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "66122a824c2ee76c5417c847", "sample_id": "66122a824c2ee76c5417c847", "file": "20_seconds/HO3D/66122a824c2ee76c5417c847_MANO_GT_20.00s.mp4", "recording_id": "SiBF11", "source_first_frame": "1479", "source_last_frame": "2078", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "66122a824c2ee76c5417c847", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/66122a824c2ee76c5417c847_MANO_GT_20.00s.mp4", "bytes": 4147584, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/66122a824c2ee76c5417c847/reference.json", "sha256": "264be2285e6aba1d8d4710c6c8b333fd33443163dd677cf55aaf9036b4647361" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "79630df9fcc743d14a94ba5fe7f029e03dd607645dd2c9e5837901765971c392", "mano_sha256": "0207725b6c785c738b000ea1195d48d01b6ec66720eaaf10abe1771298dbb98d", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "6b615ef6ef1302ad2df9f296", "sample_id": "6b615ef6ef1302ad2df9f296", "file": "20_seconds/HO3D/6b615ef6ef1302ad2df9f296_MANO_GT_20.00s.mp4", "recording_id": "SiBF10", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "6b615ef6ef1302ad2df9f296", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/6b615ef6ef1302ad2df9f296_MANO_GT_20.00s.mp4", "bytes": 4866693, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/6b615ef6ef1302ad2df9f296/reference.json", "sha256": "ed72b66833baa9f022622907fb0074ee6ca9a4c0a6955b4307dae1f516386b9a" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0f0be53d6ab9ce381a49beb7a0626dc1e3ff8f6fbffe832d01d850f41942f503", "mano_sha256": "05be4dd56375070a29f5220aa4fd429666354cc14af14e5e52de99ebc2a81080", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "6f85feb68dda0ac6365bf624", "sample_id": "6f85feb68dda0ac6365bf624", "file": "20_seconds/HO3D/6f85feb68dda0ac6365bf624_MANO_GT_20.00s.mp4", "recording_id": "GSF10", "source_first_frame": "600", "source_last_frame": "1199", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "6f85feb68dda0ac6365bf624", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/6f85feb68dda0ac6365bf624_MANO_GT_20.00s.mp4", "bytes": 5180963, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/6f85feb68dda0ac6365bf624/reference.json", "sha256": "6c6445a69deda88e41c4ef181390f69e7b76e5100ba4bd405b1b290355d0f624" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "140e57175a21ea8f312533717346dc164fe04f2ce12c10444d7cadb28f33902b", "mano_sha256": "e5e28054385e6185b2eda3c0addd89943aa5065535657ed7a058d3849cd0131e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "731f0802782b6934f3a3db56", "sample_id": "731f0802782b6934f3a3db56", "file": "20_seconds/HO3D/731f0802782b6934f3a3db56_MANO_GT_20.00s.mp4", "recording_id": "SM5", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "731f0802782b6934f3a3db56", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/731f0802782b6934f3a3db56_MANO_GT_20.00s.mp4", "bytes": 5544548, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/731f0802782b6934f3a3db56/reference.json", "sha256": "9987631e514725b34b0cb13d35db74d8f2b5beeff871cee97a7786a6477375ce" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5e3dc89fb2cc69a78703eceae4e826dec0015de7997c6b71fd664fb8a68bee71", "mano_sha256": "4b5c6abb6d71b2166312b494a00b4f6253b49d89c281e1b03ca3492c8fd36499", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "80a1ef6f1b7938153a3de410", "sample_id": "80a1ef6f1b7938153a3de410", "file": "20_seconds/HO3D/80a1ef6f1b7938153a3de410_MANO_GT_20.00s.mp4", "recording_id": "GSF13", "source_first_frame": "600", "source_last_frame": "1199", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "80a1ef6f1b7938153a3de410", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/80a1ef6f1b7938153a3de410_MANO_GT_20.00s.mp4", "bytes": 7449469, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/80a1ef6f1b7938153a3de410/reference.json", "sha256": "8841a3c3de6eb671a7dd54fd6635904888b771768a5a1e2da7fbc1b7c76c17fb" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ccff4ce1eceeeb77d0d36a6dcb3fb39e063ee96e9f9f34f50180ad56bf706162", "mano_sha256": "84abe99f08dbc3604d6255f70ecda36cd983e794dd9d375e7b25e44f6d19f561", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "8414670c57cfe47b235ed26b", "sample_id": "8414670c57cfe47b235ed26b", "file": "20_seconds/HO3D/8414670c57cfe47b235ed26b_MANO_GT_20.00s.mp4", "recording_id": "SM2", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "8414670c57cfe47b235ed26b", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/8414670c57cfe47b235ed26b_MANO_GT_20.00s.mp4", "bytes": 5850281, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/8414670c57cfe47b235ed26b/reference.json", "sha256": "f7f48a546a12721ad3e84e6fcf05b8194059b01a4936e007f652455a1cc1548a" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "fcbeb0b8badc38af84f8212e4fef3574bf468982389af0c25663c6a4362ed128", "mano_sha256": "ac04f164acaa62b3193e1e33e8feb2be305aff2599e703d861df2de1226b9f49", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "860b42f3631125e457a21c50", "sample_id": "860b42f3631125e457a21c50", "file": "20_seconds/HO3D/860b42f3631125e457a21c50_MANO_GT_20.00s.mp4", "recording_id": "SiBF12", "source_first_frame": "1479", "source_last_frame": "2078", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "860b42f3631125e457a21c50", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/860b42f3631125e457a21c50_MANO_GT_20.00s.mp4", "bytes": 4462960, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/860b42f3631125e457a21c50/reference.json", "sha256": "faf50f1d91b4e48688788c374b585d0eba11d8fa0fea8e328d507581d412b380" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f4cb7ae01763eac751ee60ee671d150c76e4550d210517962e54b088dfa7446c", "mano_sha256": "6b1d344e30cee6c0e4b5698aaa84f8797c98cb04a03d1769cfa3fb06a37c58ba", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "87585371cb0654a35301428a", "sample_id": "87585371cb0654a35301428a", "file": "20_seconds/HO3D/87585371cb0654a35301428a_MANO_GT_20.00s.mp4", "recording_id": "SMu1", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "87585371cb0654a35301428a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/87585371cb0654a35301428a_MANO_GT_20.00s.mp4", "bytes": 3648847, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/87585371cb0654a35301428a/reference.json", "sha256": "547475169835298faa37360813477a0122b406d501c0b662ba7cb66b7123c943" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "9578a6ce961a64b1fffa914c37976b933b441b9f7a47b46aa8b554b7e82bd9be", "mano_sha256": "1b88dcab78cb6edb7ca2640beddbe44635f5f8b94a19bc9d1e8a0e2fa7551958", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "9401f7a3c89872fe60c7858e", "sample_id": "9401f7a3c89872fe60c7858e", "file": "20_seconds/HO3D/9401f7a3c89872fe60c7858e_MANO_GT_20.00s.mp4", "recording_id": "BB14", "source_first_frame": "413", "source_last_frame": "1012", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "9401f7a3c89872fe60c7858e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/9401f7a3c89872fe60c7858e_MANO_GT_20.00s.mp4", "bytes": 4408768, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/9401f7a3c89872fe60c7858e/reference.json", "sha256": "6b92a8ed117217b4e481e1e4ffdcb39c614cf8b2613c61d86114d553cf9ec821" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3ce4c0fcdd9ae05ca3b3b575fc7cfae9ddf987def025f28ce1059b250df6085d", "mano_sha256": "7f2ae6473b00ccd3fbe628bf0ce5996fbc1b4bb359a124bc505374716aefcb8e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "996df080a29d4f861046ead5", "sample_id": "996df080a29d4f861046ead5", "file": "20_seconds/HO3D/996df080a29d4f861046ead5_MANO_GT_20.00s.mp4", "recording_id": "SMu42", "source_first_frame": "600", "source_last_frame": "1199", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "996df080a29d4f861046ead5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/996df080a29d4f861046ead5_MANO_GT_20.00s.mp4", "bytes": 4523000, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/996df080a29d4f861046ead5/reference.json", "sha256": "d78270cee1d9d439737736088357b4aeb315df4d1d21cfee516aecd5bcacf76f" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "608a3c3a9b6b65adc32509a894a0a1aab9c9fe58792c578d4dca39e942d777b5", "mano_sha256": "13244958123806d04e214319989095753a0fb3ae5d1ae8adc5e8039d0013a517", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "a4cceb0d7be37c326d6652a1", "sample_id": "a4cceb0d7be37c326d6652a1", "file": "20_seconds/HO3D/a4cceb0d7be37c326d6652a1_MANO_GT_20.00s.mp4", "recording_id": "GSF12", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "a4cceb0d7be37c326d6652a1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/a4cceb0d7be37c326d6652a1_MANO_GT_20.00s.mp4", "bytes": 5475036, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/a4cceb0d7be37c326d6652a1/reference.json", "sha256": "3f18e490c5a3bee457c064a3b1d928a94b6a603a39cc85ef413dce105190a6d3" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "dca8d28204b5a351232f47175429535cd27c55af6a3b8c0e69f989d41235e868", "mano_sha256": "cf93213008e8feb5b0ecd8acc7d82d26b85daa41060b3b619518bd2ed841f4af", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "a799a6a01330d0b59dafe81f", "sample_id": "a799a6a01330d0b59dafe81f", "file": "20_seconds/HO3D/a799a6a01330d0b59dafe81f_MANO_GT_20.00s.mp4", "recording_id": "SMu41", "source_first_frame": "1200", "source_last_frame": "1799", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "a799a6a01330d0b59dafe81f", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/a799a6a01330d0b59dafe81f_MANO_GT_20.00s.mp4", "bytes": 4081027, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/a799a6a01330d0b59dafe81f/reference.json", "sha256": "9281498c0b2a0eaae243e07a0aef342c421b1e00ec1afe2a38fa4200db538e68" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c9f0af038d0f7ce488aab471c2e0e71654885066f82be74587ac5add2bd6107f", "mano_sha256": "daa12fa1d880910355f4c602e26fb50e2d03473b5d22378427a1e717d7250712", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "adc4fc2c8685bf4466b8f37c", "sample_id": "adc4fc2c8685bf4466b8f37c", "file": "20_seconds/HO3D/adc4fc2c8685bf4466b8f37c_MANO_GT_20.00s.mp4", "recording_id": "MC6", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "adc4fc2c8685bf4466b8f37c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/adc4fc2c8685bf4466b8f37c_MANO_GT_20.00s.mp4", "bytes": 6654031, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/adc4fc2c8685bf4466b8f37c/reference.json", "sha256": "f2b5c0d47fd68a430db0d919d54a42609088783ce23f36c38a78317cddba7fe7" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0f3cfac17ef201edc9216dd05de34b26119bb37d47271fcdc9146431af03064a", "mano_sha256": "35a13aff44fff47a535a61395236ae507f42ceb31689dde21b00591f61c90209", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "bba2095ef3af7a85d35ec027", "sample_id": "bba2095ef3af7a85d35ec027", "file": "20_seconds/HO3D/bba2095ef3af7a85d35ec027_MANO_GT_20.00s.mp4", "recording_id": "SM3", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "bba2095ef3af7a85d35ec027", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/bba2095ef3af7a85d35ec027_MANO_GT_20.00s.mp4", "bytes": 5649690, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/bba2095ef3af7a85d35ec027/reference.json", "sha256": "52ce3a2a888be7fc2c9c3bd29b5bb9ded3297e45dc17adde487f38ac3fa56243" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b6b6ab9a9ee0927b0f1535ac1ada4eb7cf6fb73f95f26fcbbe12e24feca1f06e", "mano_sha256": "c7744574d18539acac71d4073438ef63fa5361344bd2f2d4caba0aa4b9c44691", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "d27988e7753a148e46b943a3", "sample_id": "d27988e7753a148e46b943a3", "file": "20_seconds/HO3D/d27988e7753a148e46b943a3_MANO_GT_20.00s.mp4", "recording_id": "SMu40", "source_first_frame": "1200", "source_last_frame": "1799", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "d27988e7753a148e46b943a3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/d27988e7753a148e46b943a3_MANO_GT_20.00s.mp4", "bytes": 4908206, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/d27988e7753a148e46b943a3/reference.json", "sha256": "32363e938525b900694eddfde28a6fbdf8e06244165f2350a336b6834beb5dec" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a8cdf05c2c27976d67074ed69d0308885c909eb9f4ab86831ed577fe1385a18a", "mano_sha256": "bafe0831c8654c78b463c59748732779b0e2464c6161d6f7efa1523010ccf914", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth; missing regions completed with sensor-scale-aligned predicted DA3 depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "d3ae001488076656eccfb4a9", "sample_id": "d3ae001488076656eccfb4a9", "file": "20_seconds/HO3D/d3ae001488076656eccfb4a9_MANO_GT_20.00s.mp4", "recording_id": "ND2", "source_first_frame": "703", "source_last_frame": "1302", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "d3ae001488076656eccfb4a9", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/d3ae001488076656eccfb4a9_MANO_GT_20.00s.mp4", "bytes": 4290080, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度＋管线补全", "kind": "mixed", "detail": "实测深度融合，并用按传感器尺度对齐的 DA3 预测深度补齐缺失区域；补全部分不是 GT。" }, "camera": { "label": "固定相机原点", "kind": "coordinate", "detail": "该片段没有原生外参，以固定相机光心作为世界坐标原点；只是坐标系约定，不是估计出的轨迹或外部位置 GT。", "evidence": { "file": "long20_20260918/local/extra/previews/d3ae001488076656eccfb4a9/reference.json", "sha256": "f7fde5e785a58da140ca8733556c6167208bcb1688d20cb9a45a0e6c7d748b76" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "600fe48d8741035d18a44b98c04d876e198e0673fe993f27ffb1469577b296de", "mano_sha256": "063d95fe8bb832b030c8c510b0150ac8fc9b0f1241e039541adca300b8b9add6", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "d71da6f29fa78d697d545c73", "sample_id": "d71da6f29fa78d697d545c73", "file": "20_seconds/HO3D/d71da6f29fa78d697d545c73_MANO_GT_20.00s.mp4", "recording_id": "GSF10", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "d71da6f29fa78d697d545c73", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/d71da6f29fa78d697d545c73_MANO_GT_20.00s.mp4", "bytes": 5060944, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/d71da6f29fa78d697d545c73/reference.json", "sha256": "f6f981f53aea8045c77e3a991c981f75dd4261141404e6928a084c7a71c1f8b7" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4af5cd54ecd42f5ec5139c3eb887647eb9bdc2f55bdc28ca62e5a3df0268fd28", "mano_sha256": "e2ca1bacc0cb7d18115b99a34524366b8eaaa0a63da3bbb6e677bc66f92ca3f2", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "d7e89e943f5948b9b85074f5", "sample_id": "d7e89e943f5948b9b85074f5", "file": "20_seconds/HO3D/d7e89e943f5948b9b85074f5_MANO_GT_20.00s.mp4", "recording_id": "SiBF11", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "d7e89e943f5948b9b85074f5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/d7e89e943f5948b9b85074f5_MANO_GT_20.00s.mp4", "bytes": 4462825, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/d7e89e943f5948b9b85074f5/reference.json", "sha256": "f444d2a77bfeb781181e6bab176ad205ea65a745d121bc7e2d96b003e743deb1" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "61e69c5392d0cac355f55f13f5788536dd1ad0c414c7327212169f31b9058b7f", "mano_sha256": "fab8247fe16dbd4cb7f1149476b65b31181f5f7740748cccc5f18f863485031d", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "d9faac5a25b2552f8fd82203", "sample_id": "d9faac5a25b2552f8fd82203", "file": "20_seconds/HO3D/d9faac5a25b2552f8fd82203_MANO_GT_20.00s.mp4", "recording_id": "SB10", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "d9faac5a25b2552f8fd82203", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/d9faac5a25b2552f8fd82203_MANO_GT_20.00s.mp4", "bytes": 5264431, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/d9faac5a25b2552f8fd82203/reference.json", "sha256": "3576a675049986242ab06f37e28843b3c9a82061674a8afed2b6d0c37b10c376" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ba6f9ee3d193a9be99b8825e1602488141e588b71a66fc0c2a6c3c086a54ee58", "mano_sha256": "adaaddb06a7a99e8b9ecd1969bde9ef9e8dbe05f67992502d8a70887b9eb70b6", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "dab423e4d0b746cf86bb73e3", "sample_id": "dab423e4d0b746cf86bb73e3", "file": "20_seconds/HO3D/dab423e4d0b746cf86bb73e3_MANO_GT_20.00s.mp4", "recording_id": "BB11", "source_first_frame": "413", "source_last_frame": "1012", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "dab423e4d0b746cf86bb73e3", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/dab423e4d0b746cf86bb73e3_MANO_GT_20.00s.mp4", "bytes": 4180581, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/dab423e4d0b746cf86bb73e3/reference.json", "sha256": "7d66d7bd6300fb5cb777b0eabe72036059064c5cfc9d275440bb24bcbc0901ce" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f368c65391fc4dcbd78c1e3c5c3adce6cf4114db7d7b52c14ad40dbe6d08874f", "mano_sha256": "6471126a0956d854096a49ebe3c1b54ba21ec0e52910e8e238039375a3129b05", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "e2287df90ed8977e190861a0", "sample_id": "e2287df90ed8977e190861a0", "file": "20_seconds/HO3D/e2287df90ed8977e190861a0_MANO_GT_20.00s.mp4", "recording_id": "BB12", "source_first_frame": "413", "source_last_frame": "1012", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "e2287df90ed8977e190861a0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/e2287df90ed8977e190861a0_MANO_GT_20.00s.mp4", "bytes": 4256852, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/e2287df90ed8977e190861a0/reference.json", "sha256": "9224a285088173102ce29c46f322f4bf1eb53c22328324418b62e6d617927bc9" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "591d7f75f8cfaaeca3f6209b17054b7f6cb618efa81219f162d8088c16838381", "mano_sha256": "cd26a979590688070fc9d6e9a0ce248a0a5a4d4b2dd19f19c94f9bf337ebb47a", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "e3bf3b45ac80ecddcd47297c", "sample_id": "e3bf3b45ac80ecddcd47297c", "file": "20_seconds/HO3D/e3bf3b45ac80ecddcd47297c_MANO_GT_20.00s.mp4", "recording_id": "SMu40", "source_first_frame": "600", "source_last_frame": "1199", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "e3bf3b45ac80ecddcd47297c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/e3bf3b45ac80ecddcd47297c_MANO_GT_20.00s.mp4", "bytes": 4735656, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/e3bf3b45ac80ecddcd47297c/reference.json", "sha256": "52598ae7e0c3979f0af047799da6131eb5156153ae105c44214e61dd19f2f705" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "c793c3fb4f9ee1b2471d38914cfc70e9836eb250e7c0052939ee35c78ac762af", "mano_sha256": "d2184ec1ae00aba97da5fe1798028c9bf185f1c9754b91972656210e5fc9b53f", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "e68ca76d73d04f4e5f1b8790", "sample_id": "e68ca76d73d04f4e5f1b8790", "file": "20_seconds/HO3D/e68ca76d73d04f4e5f1b8790_MANO_GT_20.00s.mp4", "recording_id": "GSF11", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "e68ca76d73d04f4e5f1b8790", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/e68ca76d73d04f4e5f1b8790_MANO_GT_20.00s.mp4", "bytes": 4535317, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/e68ca76d73d04f4e5f1b8790/reference.json", "sha256": "66f17221c82c1dee9d7a768fb6b9f58134b324b04c5a7111a2f14514384c77d5" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "534911d5af70912609e922cfb07c9a51b8ba35869da9244411fa369209f983b4", "mano_sha256": "e3efc88431c24570e219231d3ed609ff190c40fb73afaf9b5f1c3661e307544e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "e9a1876df01fe028ca7ed967", "sample_id": "e9a1876df01fe028ca7ed967", "file": "20_seconds/HO3D/e9a1876df01fe028ca7ed967_MANO_GT_20.00s.mp4", "recording_id": "SMu42", "source_first_frame": "1200", "source_last_frame": "1799", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "e9a1876df01fe028ca7ed967", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/e9a1876df01fe028ca7ed967_MANO_GT_20.00s.mp4", "bytes": 4797219, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/e9a1876df01fe028ca7ed967/reference.json", "sha256": "53d7406e1591c813bb3a8b0f6801f05bbc9ba4ad50eb37dc589f2e5c042e1660" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e66fb744931c1f8952d2bc3babc6d24f6068ad7b744b6fd6204cc1286438aae4", "mano_sha256": "b4e827e26b671259b82b6c380495df3390b2e1186fefb3a2f0a1720026799980", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "ef1af3542d3e0e9cb52ff2f8", "sample_id": "ef1af3542d3e0e9cb52ff2f8", "file": "20_seconds/HO3D/ef1af3542d3e0e9cb52ff2f8_MANO_GT_20.00s.mp4", "recording_id": "ShSu13", "source_first_frame": "21", "source_last_frame": "620", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "ef1af3542d3e0e9cb52ff2f8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/ef1af3542d3e0e9cb52ff2f8_MANO_GT_20.00s.mp4", "bytes": 5171913, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/ef1af3542d3e0e9cb52ff2f8/reference.json", "sha256": "919705417d0b68f83d3dbba3fa992799276c6b091161c003a1f7f44c64dafafd" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "708f6ff5ff24d6d73a8ff3ad56f8a9bfc4963e13596cd4111927ed59b9e115e4", "mano_sha256": "47cefefde5f98fee040b99e8a50e23d5fe13e4fa7cb763fac603a7b8c1fd71ec", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "f316a3b6d2862719d1a28952", "sample_id": "f316a3b6d2862719d1a28952", "file": "20_seconds/HO3D/f316a3b6d2862719d1a28952_MANO_GT_20.00s.mp4", "recording_id": "BB13", "source_first_frame": "413", "source_last_frame": "1012", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "f316a3b6d2862719d1a28952", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/f316a3b6d2862719d1a28952_MANO_GT_20.00s.mp4", "bytes": 4115636, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/f316a3b6d2862719d1a28952/reference.json", "sha256": "cf90543f077a9cd4d7a6e9103036ee75a85ee24709b66942b0b6ef88912c673e" } } } }, { "dataset": "HO3D", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "001f49dcfcf2976c33fdaf126e33997b37803f2c669c6de6f947a70228aeb599", "mano_sha256": "e5c4b7ddfa688a18cadb7b188c0abdb8915567ef7233687dd210dafd69449c9e", "duration_s": 20.0, "frames": 600, "fps": 30.0, "mano_source": "native HO3D v3 MANO; official fitted annotation", "scene_depth_source": "native HO3D v3 sensor depth", "scene_camera_source": "fixed calibrated capture camera; no fabricated motion", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "f36488d99bff61c560e71efe", "sample_id": "f36488d99bff61c560e71efe", "file": "20_seconds/HO3D/f36488d99bff61c560e71efe_MANO_GT_20.00s.mp4", "recording_id": "SMu41", "source_first_frame": "0", "source_last_frame": "599", "source_split": "official_train", "training_label_source": "unavailable", "quality_warnings": "", "id": "f36488d99bff61c560e71efe", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/20_seconds/HO3D/f36488d99bff61c560e71efe_MANO_GT_20.00s.mp4", "bytes": 4019946, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "status": "unavailable", "camera_class": "unknown", "hand_class": "unknown", "relative_class": "unknown", "reason": "现有参考仅有相机系手部；世界系、有效性与时间来源尚未完成逐样本核验。" }, "camera_display_source": "使用该片段的官方固定相机外参。", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用该片段的官方固定相机外参。", "evidence": { "file": "long20_20260918/local/extra/previews/f36488d99bff61c560e71efe/reference.json", "sha256": "98793bf0ca0f45e9e3c2e49cca7a11a4c4c5cf4f708184c70b3022d07b5741a7" } } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "9b4b0d75209374f7cd64498e0d39e93c65b1711adf57d928de2c5d503cfc896c", "mano_sha256": "d4681f8bfd1c7820d6a9c38cc0b0e62dda05a17145a4f8d9178134c8bc51884f", "duration_s": 2.334, "frames": 70, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "031a1118f76ad8002c8ad10a", "sample_id": "031a1118f76ad8002c8ad10a", "file": "short_originals_under_20s/DexYCB/031a1118f76ad8002c8ad10a_MANO_GT_2.33s.mp4", "recording_id": "20200820-subject-03/20200820_143043", "source_first_frame": "0", "source_last_frame": "69", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 1/70 output frames; may include genuine out-of-view frames", "id": "031a1118f76ad8002c8ad10a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/031a1118f76ad8002c8ad10a_MANO_GT_2.33s.mp4", "bytes": 696317, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 69, "valid_duration_s": 2.3, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 69, "valid_duration_s": 2.3, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.23236167072499608, "p90": 0.8396877302779693, "valid_pairs": 68, "valid_duration_s": 2.266666666666666, "coverage": 0.9855072463768114 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 68, "valid_duration_s": 2.266666666666666, "coverage": 0.9855072463768114 }, "rhcm_continuous_estimate_m_s": { "median": 0.23236167072499608, "p90": 0.8396877302779693, "valid_pairs": 68, "valid_duration_s": 2.266666666666666, "coverage": 0.9855072463768114 }, "rhcm_discrete_m_s": { "median": 0.23236167072499647, "p90": 0.8396877302779704, "valid_pairs": 68, "valid_duration_s": 2.266666666666666, "coverage": 0.9855072463768114 }, "wms": { "median": 0.9967985864454261, "p90": 0.9988184856419727, "valid_pairs": 63, "valid_duration_s": 2.0999999999999996, "coverage": 0.9130434782608695 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 69, "hand_world_speed_p90_m_s": 0.8396877302779693, "hand_camera_speed_p90_m_s": 0.8396877302779704, "hand_world_valid_fraction": 0.4928571428571429, "hand_camera_valid_fraction": 0.4928571428571429, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.8396877302779704, "rham_p90_deg_s": null, "wms_p90": 0.9988184856419727, "gt_coverage": 0.9855072463768114, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "d093c4307b6dec7e86c52bad964fe18c872a52ff182c8b459d96e912532680bb", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/031a1118f76ad8002c8ad10a/reference.npz", "window_s": [ 0.0, 2.3 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 70, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "29f7d30cd681730a5c9478c134f7bab10d4c3541aa7b60f5b396159c5573ff97", "mano_sha256": "66889370ed76b061f9fe257bf3d6dcccd7947f527f791ea74ed74e83e2743108", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "11e2555c17c6a8f9c29f3950", "sample_id": "11e2555c17c6a8f9c29f3950", "file": "short_originals_under_20s/DexYCB/11e2555c17c6a8f9c29f3950_MANO_GT_2.47s.mp4", "recording_id": "20200928-subject-07/20200928_144300", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 13/74 output frames; may include genuine out-of-view frames", "id": "11e2555c17c6a8f9c29f3950", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/11e2555c17c6a8f9c29f3950_MANO_GT_2.47s.mp4", "bytes": 836199, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 3.122441499087811e-13, "p90": 3.1224414990878215e-13, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.18747062679126128, "p90": 0.9472297667862015, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.821917808219178 }, "cmh_m_s": { "median": 5.094566796263958e-15, "p90": 6.169013532957413e-15, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.821917808219178 }, "rhcm_continuous_estimate_m_s": { "median": 0.18747062679125892, "p90": 0.9472297667861992, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.821917808219178 }, "rhcm_discrete_m_s": { "median": 0.18747061115657354, "p90": 0.9472296913700229, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.821917808219178 }, "wms": { "median": 0.994694129813718, "p90": 0.9989453718575555, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.821917808219178 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 3.1224414990878215e-13, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.9472297667862015, "hand_camera_speed_p90_m_s": 0.9472296913700229, "hand_world_valid_fraction": 0.41216216216216217, "hand_camera_valid_fraction": 0.41216216216216217, "cmh_p90_m_s": 6.169013532957413e-15, "rhcm_p90_m_s": 0.9472296913700229, "rham_p90_deg_s": null, "wms_p90": 0.9989453718575555, "gt_coverage": 0.821917808219178, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "f3c7f2268769eafbd052303c369f4535f30e3c781af392dc5dd351dc4c8ca268", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/11e2555c17c6a8f9c29f3950/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f8f7a2d4ef0269ca75018222a31578103fb5687f4d95791c21b0894cea03b411", "mano_sha256": "8eab76f0060fcd3df62f431bc22022213521abd0c53c8d77edde3fa751c1b9ba", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "1275f01ebb260ad8efdacd79", "sample_id": "1275f01ebb260ad8efdacd79", "file": "short_originals_under_20s/DexYCB/1275f01ebb260ad8efdacd79_MANO_GT_2.40s.mp4", "recording_id": "20200903-subject-04/20200903_113634", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 5/72 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 3 frames", "id": "1275f01ebb260ad8efdacd79", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/1275f01ebb260ad8efdacd79_MANO_GT_2.40s.mp4", "bytes": 751224, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.4522728325182757, "p90": 0.9647608298802574, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "rhcm_continuous_estimate_m_s": { "median": 0.4522728325182757, "p90": 0.9647608298802574, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "rhcm_discrete_m_s": { "median": 0.4522728325182752, "p90": 0.9647608298802575, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "wms": { "median": 0.9977934785078055, "p90": 0.9989644352661214, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.9647608298802574, "hand_camera_speed_p90_m_s": 0.9647608298802575, "hand_world_valid_fraction": 0.4652777777777778, "hand_camera_valid_fraction": 0.4652777777777778, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.9647608298802575, "rham_p90_deg_s": null, "wms_p90": 0.9989644352661214, "gt_coverage": 0.9295774647887323, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "dfdd68ef018f6a964fa1f8a427e9de3174c5df7ccb05436a5e84693a35fbe641", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/1275f01ebb260ad8efdacd79/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "586eccc07a8243e6d9c3a3e08bf7bd03132f0cee151f3346fdcc9a7f9cf59eb2", "mano_sha256": "4c6835d34eadbc552b6af609683364b6e5cd88d14a3cfc42453354e10c6ed866", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "24280961cedbe892cb340365", "sample_id": "24280961cedbe892cb340365", "file": "short_originals_under_20s/DexYCB/24280961cedbe892cb340365_MANO_GT_2.40s.mp4", "recording_id": "20200709-subject-01/20200709_150654", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 5/72 output frames; may include genuine out-of-view frames", "id": "24280961cedbe892cb340365", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/24280961cedbe892cb340365_MANO_GT_2.40s.mp4", "bytes": 782866, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.19961001854547022, "p90": 1.0766900322610071, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "rhcm_continuous_estimate_m_s": { "median": 0.19961001854547022, "p90": 1.0766900322610071, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "rhcm_discrete_m_s": { "median": 0.19961001854547067, "p90": 1.0766900322610065, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "wms": { "median": 0.9950128702268496, "p90": 0.9990720894011529, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9295774647887323 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0766900322610071, "hand_camera_speed_p90_m_s": 1.0766900322610065, "hand_world_valid_fraction": 0.4652777777777778, "hand_camera_valid_fraction": 0.4652777777777778, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0766900322610065, "rham_p90_deg_s": null, "wms_p90": 0.9990720894011529, "gt_coverage": 0.9295774647887323, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "f2a184c8e148ad338d3702b83689451316553451575e263461351d53b4494c35", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/24280961cedbe892cb340365/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7ebd18e55b84fe8fb129ed9b649a99939641f409a8e62ae6fec1fa5027da4b1c", "mano_sha256": "3cba478562c56269ddb15c1d2b4e02513ea30ae5c1e5b48dd53971d9d5f52041", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "30159c46b15efbc6f3c992d6", "sample_id": "30159c46b15efbc6f3c992d6", "file": "short_originals_under_20s/DexYCB/30159c46b15efbc6f3c992d6_MANO_GT_2.47s.mp4", "recording_id": "20200928-subject-07/20200928_145626", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 15/74 output frames; may include genuine out-of-view frames", "id": "30159c46b15efbc6f3c992d6", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/30159c46b15efbc6f3c992d6_MANO_GT_2.47s.mp4", "bytes": 864053, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 3.9843289069162666e-13, "p90": 3.98432890691628e-13, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.20902599156110163, "p90": 0.718100255202436, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "cmh_m_s": { "median": 5.498480039267321e-15, "p90": 5.657641906952465e-15, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "rhcm_continuous_estimate_m_s": { "median": 0.2090259915611058, "p90": 0.7181002552024385, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "rhcm_discrete_m_s": { "median": 0.20902597018974287, "p90": 0.7181001836464631, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "wms": { "median": 0.99523672302487, "p90": 0.9985940218957795, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 3.98432890691628e-13, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.718100255202436, "hand_camera_speed_p90_m_s": 0.7181001836464631, "hand_world_valid_fraction": 0.39864864864864863, "hand_camera_valid_fraction": 0.39864864864864863, "cmh_p90_m_s": 5.657641906952465e-15, "rhcm_p90_m_s": 0.7181001836464631, "rham_p90_deg_s": null, "wms_p90": 0.9985940218957795, "gt_coverage": 0.7945205479452054, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "931422bcb7dca08b8bcedea72a176d284b7bd1787b121a6378cc0c2c2ac90d70", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/30159c46b15efbc6f3c992d6/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "cdd8a661aae6588b090afeee82744ef8e75073e3bd4fbfe4866e3530ea946f85", "mano_sha256": "4dbf76e1f198cd19aabe4d1cafe563d54f941e88141faf4034377d63f885e8ac", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "3113f8c9903c7690d04b612e", "sample_id": "3113f8c9903c7690d04b612e", "file": "short_originals_under_20s/DexYCB/3113f8c9903c7690d04b612e_MANO_GT_2.47s.mp4", "recording_id": "20200813-subject-02/20200813_150535", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_validation", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "valid mesh not rasterized in 6 frames", "id": "3113f8c9903c7690d04b612e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/3113f8c9903c7690d04b612e_MANO_GT_2.47s.mp4", "bytes": 878435, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.3177307656550032, "p90": 1.1593420970929873, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.3177307656550032, "p90": 1.1593420970929873, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.3177307656550032, "p90": 1.1593420970929873, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "wms": { "median": 0.9975820791619726, "p90": 0.9991492737838426, "valid_pairs": 62, "valid_duration_s": 2.066666666666667, "coverage": 0.8493150684931509 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 1.1593420970929873, "hand_camera_speed_p90_m_s": 1.1593420970929873, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.1593420970929873, "rham_p90_deg_s": null, "wms_p90": 0.9991492737838426, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "a5962979824223f4fc7af9ecba7193aa194391befc2f9c9161e3566267007fd5", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/3113f8c9903c7690d04b612e/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6bb7dbd4362c056940546b1e332e6264b634c99c4a32605bca97dbd26e1c086b", "mano_sha256": "4b91a1cd4d4f9f488c7e53870b367adec4bc479d00b9901f72243ba6e5bba43b", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "31b74c078b206002de08ac98", "sample_id": "31b74c078b206002de08ac98", "file": "short_originals_under_20s/DexYCB/31b74c078b206002de08ac98_MANO_GT_2.40s.mp4", "recording_id": "20201015-subject-09/20201015_150506", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "valid mesh not rasterized in 16 frames", "id": "31b74c078b206002de08ac98", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/31b74c078b206002de08ac98_MANO_GT_2.40s.mp4", "bytes": 825306, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.05847422143370564, "p90": 0.7523174157816114, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.05847422143370564, "p90": 0.7523174157816114, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.0584742214337051, "p90": 0.752317415781611, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "wms": { "median": 0.992628629540599, "p90": 0.9986892648790717, "valid_pairs": 54, "valid_duration_s": 1.8, "coverage": 0.7605633802816901 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.7523174157816114, "hand_camera_speed_p90_m_s": 0.752317415781611, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.752317415781611, "rham_p90_deg_s": null, "wms_p90": 0.9986892648790717, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "fda24a92d1a1e60eede17441339b62ee00fdb9fc81f5e6e58f47d49755beec52", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/31b74c078b206002de08ac98/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "058ca5710e8e6e620b7c8032fd7ed208e8e55eb836cfbd350f724ed219f431d9", "mano_sha256": "1a3c5e1e805b3cb0701be2981b4afe7105082b641314e26678c41feb56ddab84", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "32014d197dd6763559e940c0", "sample_id": "32014d197dd6763559e940c0", "file": "short_originals_under_20s/DexYCB/32014d197dd6763559e940c0_MANO_GT_2.47s.mp4", "recording_id": "20200908-subject-05/20200908_151441", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 15/74 output frames; may include genuine out-of-view frames", "id": "32014d197dd6763559e940c0", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/32014d197dd6763559e940c0_MANO_GT_2.47s.mp4", "bytes": 783806, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 1.2985338845361775e-14, "p90": 1.2985338845361818e-14, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.25685106786117773, "p90": 0.7856715412553393, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "cmh_m_s": { "median": 1.7432444400134374e-16, "p90": 2.4556288172643216e-16, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "rhcm_continuous_estimate_m_s": { "median": 0.2568510678611776, "p90": 0.7856715412553392, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "rhcm_discrete_m_s": { "median": 0.256851061441294, "p90": 0.7856715008723708, "valid_pairs": 58, "valid_duration_s": 1.9333333333333331, "coverage": 0.7945205479452054 }, "wms": { "median": 0.9963048061565483, "p90": 0.9987297828443934, "valid_pairs": 56, "valid_duration_s": 1.8666666666666665, "coverage": 0.7671232876712328 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 1.2985338845361818e-14, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.7856715412553393, "hand_camera_speed_p90_m_s": 0.7856715008723708, "hand_world_valid_fraction": 0.39864864864864863, "hand_camera_valid_fraction": 0.39864864864864863, "cmh_p90_m_s": 2.4556288172643216e-16, "rhcm_p90_m_s": 0.7856715008723708, "rham_p90_deg_s": null, "wms_p90": 0.9987297828443934, "gt_coverage": 0.7945205479452054, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "084f1215c06ac7fb1c98a3747a90b732869721548def4cf18eddf4e8f8dc3890", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/32014d197dd6763559e940c0/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f3483c85971edeb6c2b8fda9c0f61cf60e5a86eb6de65aa7021d16b4d3f7652d", "mano_sha256": "78885b5ac731fcf8ba9a5e22311042607699fe5f2b5ddebc7c799f3b3d34b32c", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "39ab03735698c92ea630e4d1", "sample_id": "39ab03735698c92ea630e4d1", "file": "short_originals_under_20s/DexYCB/39ab03735698c92ea630e4d1_MANO_GT_2.40s.mp4", "recording_id": "20201022-subject-10/20201022_110843", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 7/72 output frames; may include genuine out-of-view frames", "id": "39ab03735698c92ea630e4d1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/39ab03735698c92ea630e4d1_MANO_GT_2.40s.mp4", "bytes": 758574, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.3272587744814337, "p90": 1.0745677985565363, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rhcm_continuous_estimate_m_s": { "median": 0.3272587744814337, "p90": 1.0745677985565363, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rhcm_discrete_m_s": { "median": 0.32725877448143464, "p90": 1.0745677985565376, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "wms": { "median": 0.996953292014128, "p90": 0.9990702206838978, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0745677985565363, "hand_camera_speed_p90_m_s": 1.0745677985565376, "hand_world_valid_fraction": 0.4513888888888889, "hand_camera_valid_fraction": 0.4513888888888889, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0745677985565376, "rham_p90_deg_s": null, "wms_p90": 0.9990702206838978, "gt_coverage": 0.9014084507042254, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "f362b1560580e98c3bc0dc93a362a79567ea4e21b403149757a46c22de98ccfd", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/39ab03735698c92ea630e4d1/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "50dd37016b6fe8deee79da0b7e12e7ae48cf51982451ba1035d7865eac2a5f3a", "mano_sha256": "4638eb60aac7d3dd75af596df14d2592386ca431cc797d21c9cb2dccc09718e0", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "3bc05d29f33409048057be79", "sample_id": "3bc05d29f33409048057be79", "file": "short_originals_under_20s/DexYCB/3bc05d29f33409048057be79_MANO_GT_2.40s.mp4", "recording_id": "20200813-subject-02/20200813_154335", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 7/72 output frames; may include genuine out-of-view frames", "id": "3bc05d29f33409048057be79", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/3bc05d29f33409048057be79_MANO_GT_2.40s.mp4", "bytes": 828330, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.29665059666045845, "p90": 1.1432102130999444, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rhcm_continuous_estimate_m_s": { "median": 0.29665059666045845, "p90": 1.1432102130999444, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rhcm_discrete_m_s": { "median": 0.29665059666045934, "p90": 1.1432102130999446, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "wms": { "median": 0.9966402806224545, "p90": 0.9991260321305242, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.1432102130999444, "hand_camera_speed_p90_m_s": 1.1432102130999446, "hand_world_valid_fraction": 0.4513888888888889, "hand_camera_valid_fraction": 0.4513888888888889, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.1432102130999446, "rham_p90_deg_s": null, "wms_p90": 0.9991260321305242, "gt_coverage": 0.9014084507042254, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "914286a6f6eba6d97affc5d3e1c03b220cc5b2dad038d8296d195958be798ee3", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/3bc05d29f33409048057be79/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "d7493fd86bf97d8ab1e3b2b70c1430300f25c249713740251c4481a9babd92ba", "mano_sha256": "4a0fc25aba8c95618e1a15ceee9db8e5b615dbf4ce1854e442fd8e49ef9e4d8c", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "42791a60635dd726c84e25e8", "sample_id": "42791a60635dd726c84e25e8", "file": "short_originals_under_20s/DexYCB/42791a60635dd726c84e25e8_MANO_GT_2.40s.mp4", "recording_id": "20200709-subject-01/20200709_150524", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_validation", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 3/72 output frames; may include genuine out-of-view frames", "id": "42791a60635dd726c84e25e8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/42791a60635dd726c84e25e8_MANO_GT_2.40s.mp4", "bytes": 920112, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.3910947894722621, "p90": 1.00268292007512, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "rhcm_continuous_estimate_m_s": { "median": 0.3910947894722621, "p90": 1.00268292007512, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "rhcm_discrete_m_s": { "median": 0.39109478947226095, "p90": 1.002682920075121, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "wms": { "median": 0.9974495659348686, "p90": 0.9990036691906935, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.00268292007512, "hand_camera_speed_p90_m_s": 1.002682920075121, "hand_world_valid_fraction": 0.4791666666666667, "hand_camera_valid_fraction": 0.4791666666666667, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.002682920075121, "rham_p90_deg_s": null, "wms_p90": 0.9990036691906935, "gt_coverage": 0.9577464788732394, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "c640aaf2d93ae65ed5d0f8e1d563e25d64ad89ef7c83c31344ad7cec52b2e376", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/42791a60635dd726c84e25e8/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "cfb5406415379d376bcdfbd2f1849469500998079f359a09cc5a3d4fdf9fa835", "mano_sha256": "d4044aa905fb25d0e151f65d9b3852205f18a0e18ba6be53747095750d1442f8", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "51841ff5a607984f28731172", "sample_id": "51841ff5a607984f28731172", "file": "short_originals_under_20s/DexYCB/51841ff5a607984f28731172_MANO_GT_2.40s.mp4", "recording_id": "20200928-subject-07/20200928_154023", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 11/72 output frames; may include genuine out-of-view frames", "id": "51841ff5a607984f28731172", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/51841ff5a607984f28731172_MANO_GT_2.40s.mp4", "bytes": 825239, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.13417487481838183, "p90": 1.0280006272308724, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8450704225352113 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8450704225352113 }, "rhcm_continuous_estimate_m_s": { "median": 0.13417487481838183, "p90": 1.0280006272308724, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8450704225352113 }, "rhcm_discrete_m_s": { "median": 0.13417487481838147, "p90": 1.0280006272308695, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8450704225352113 }, "wms": { "median": 0.9949448143996673, "p90": 0.9990504540519985, "valid_pairs": 48, "valid_duration_s": 1.5999999999999999, "coverage": 0.676056338028169 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0280006272308724, "hand_camera_speed_p90_m_s": 1.0280006272308695, "hand_world_valid_fraction": 0.4236111111111111, "hand_camera_valid_fraction": 0.4236111111111111, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0280006272308695, "rham_p90_deg_s": null, "wms_p90": 0.9990504540519985, "gt_coverage": 0.8450704225352113, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "c30f5045b7e0cb5228cac69a3ad33cda066f89057e521482212c3f45d5bcb1c3", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/51841ff5a607984f28731172/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "57c7eb5882e419c755a946716cc7a38e0bb467cf254fc884f994bc2d33462120", "mano_sha256": "cefc3e9be1062a8900c3492a90b28fbcf370aac3e177003228b6a95d90f62413", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "520f290a2ecc30edc4f8b35f", "sample_id": "520f290a2ecc30edc4f8b35f", "file": "short_originals_under_20s/DexYCB/520f290a2ecc30edc4f8b35f_MANO_GT_2.40s.mp4", "recording_id": "20200908-subject-05/20200908_150326", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 7/72 output frames; may include genuine out-of-view frames", "id": "520f290a2ecc30edc4f8b35f", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/520f290a2ecc30edc4f8b35f_MANO_GT_2.40s.mp4", "bytes": 849476, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 7.126582189827053e-14, "p90": 7.126582189827077e-14, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.1695582649337302, "p90": 0.8157753388314898, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "cmh_m_s": { "median": 1.106249195847114e-15, "p90": 1.1179984351000127e-15, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rhcm_continuous_estimate_m_s": { "median": 0.16955826493373094, "p90": 0.81577533883149, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rhcm_discrete_m_s": { "median": 0.16955825601762137, "p90": 0.8157751991798041, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "wms": { "median": 0.9942412286410557, "p90": 0.99878137081958, "valid_pairs": 63, "valid_duration_s": 2.1, "coverage": 0.8873239436619719 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 7.126582189827077e-14, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.8157753388314898, "hand_camera_speed_p90_m_s": 0.8157751991798041, "hand_world_valid_fraction": 0.4513888888888889, "hand_camera_valid_fraction": 0.4513888888888889, "cmh_p90_m_s": 1.1179984351000127e-15, "rhcm_p90_m_s": 0.8157751991798041, "rham_p90_deg_s": null, "wms_p90": 0.99878137081958, "gt_coverage": 0.9014084507042254, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "e2835d3a2dd196c58dbab2d111694aa82954f860340001f017a4fb5b6f089208", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/520f290a2ecc30edc4f8b35f/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "e2abe9fe5211a2adbfa4e8243026c404e2b86bfe8008152b9f72c0de9f604a2f", "mano_sha256": "47b7465325df0d3bb5ba40081c789ba91bb37b7ccaf9826fe0ab4d2bcd94bb90", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "575d67c8ad94266004323f12", "sample_id": "575d67c8ad94266004323f12", "file": "short_originals_under_20s/DexYCB/575d67c8ad94266004323f12_MANO_GT_2.40s.mp4", "recording_id": "20200813-subject-02/20200813_150456", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "", "id": "575d67c8ad94266004323f12", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/575d67c8ad94266004323f12_MANO_GT_2.40s.mp4", "bytes": 802800, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.36428484669132455, "p90": 1.091244583635677, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.36428484669132455, "p90": 1.091244583635677, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.3642848466913243, "p90": 1.0912445836356741, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "wms": { "median": 0.9974558160372468, "p90": 0.9990887749393204, "valid_pairs": 59, "valid_duration_s": 1.9666666666666661, "coverage": 0.8309859154929575 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.091244583635677, "hand_camera_speed_p90_m_s": 1.0912445836356741, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0912445836356741, "rham_p90_deg_s": null, "wms_p90": 0.9990887749393204, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "300501ac5d1b8f4d224b77ddc1ddeb64b781ed2e0a80e247e61b19d9f78b2b61", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/575d67c8ad94266004323f12/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7c147f09519c1d67e51810cd088190ff9898b5f8cfe754ef199bdeb2869b5146", "mano_sha256": "5f0e9dc751452047eedc97ce22500c823c70baa895bfe6af8dfbc84c94e35c71", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "63a6e8731a10d4d242aa43e2", "sample_id": "63a6e8731a10d4d242aa43e2", "file": "short_originals_under_20s/DexYCB/63a6e8731a10d4d242aa43e2_MANO_GT_2.40s.mp4", "recording_id": "20201015-subject-09/20201015_142844", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_test", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "", "id": "63a6e8731a10d4d242aa43e2", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/63a6e8731a10d4d242aa43e2_MANO_GT_2.40s.mp4", "bytes": 828236, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 5.219065549294893e-13, "p90": 5.21906554929491e-13, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.13945717936384452, "p90": 0.5700222625249429, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 6.476814944981392e-15, "p90": 6.7296699771020465e-15, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.13945717936383834, "p90": 0.5700222625249455, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.13945716982015793, "p90": 0.5700222403618916, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "wms": { "median": 0.9964134376967628, "p90": 0.9988569786549477, "valid_pairs": 51, "valid_duration_s": 1.7, "coverage": 0.7183098591549295 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 5.21906554929491e-13, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.5700222625249429, "hand_camera_speed_p90_m_s": 0.5700222403618916, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 6.7296699771020465e-15, "rhcm_p90_m_s": 0.5700222403618916, "rham_p90_deg_s": null, "wms_p90": 0.9988569786549477, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "3df3c831202f09fa0c5d515ca20cbd854cdef4541f2a1fd0222639cc2fd7182e", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/63a6e8731a10d4d242aa43e2/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2555a7a560a889256b678deb23b3e88760c820b30ac77b7df933eb56bcb28088", "mano_sha256": "4e2e016f228019820ee475d61bda48858173efb3934f435c726da3e5a5a3b3c7", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "65af17bb1f65429b9126eb90", "sample_id": "65af17bb1f65429b9126eb90", "file": "short_originals_under_20s/DexYCB/65af17bb1f65429b9126eb90_MANO_GT_2.40s.mp4", "recording_id": "20201002-subject-08/20201002_110741", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 19/72 output frames; may include genuine out-of-view frames", "id": "65af17bb1f65429b9126eb90", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/65af17bb1f65429b9126eb90_MANO_GT_2.40s.mp4", "bytes": 762928, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.9704597497539403e-13, "p90": 2.97045974975395e-13, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.31271244004525134, "p90": 1.0309046294908644, "valid_pairs": 52, "valid_duration_s": 1.7333333333333334, "coverage": 0.7323943661971831 }, "cmh_m_s": { "median": 3.210207014523223e-15, "p90": 4.514725217110745e-15, "valid_pairs": 52, "valid_duration_s": 1.7333333333333334, "coverage": 0.7323943661971831 }, "rhcm_continuous_estimate_m_s": { "median": 0.31271244004525267, "p90": 1.0309046294908641, "valid_pairs": 52, "valid_duration_s": 1.7333333333333334, "coverage": 0.7323943661971831 }, "rhcm_discrete_m_s": { "median": 0.312712438942034, "p90": 1.0309046577720926, "valid_pairs": 52, "valid_duration_s": 1.7333333333333334, "coverage": 0.7323943661971831 }, "wms": { "median": 0.9968036061504646, "p90": 0.9990308579091652, "valid_pairs": 52, "valid_duration_s": 1.7333333333333334, "coverage": 0.7323943661971831 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 2.97045974975395e-13, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0309046294908644, "hand_camera_speed_p90_m_s": 1.0309046577720926, "hand_world_valid_fraction": 0.3680555555555556, "hand_camera_valid_fraction": 0.3680555555555556, "cmh_p90_m_s": 4.514725217110745e-15, "rhcm_p90_m_s": 1.0309046577720926, "rham_p90_deg_s": null, "wms_p90": 0.9990308579091652, "gt_coverage": 0.7323943661971831, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "8c3e79cb203b6d480de004566aed536ab68f73b8f41aec9cabc3f264f655d4f0", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/65af17bb1f65429b9126eb90/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "184a1eda6d6f5e0697bb1e0cc54f187eb662534716eaed928230c9caf18f8c56", "mano_sha256": "af3b3cd365a85a87e9524c24af3c4e3281eb9d6cd93d4b42add29695bad30615", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "6e86f351667f5a91324257bd", "sample_id": "6e86f351667f5a91324257bd", "file": "short_originals_under_20s/DexYCB/6e86f351667f5a91324257bd_MANO_GT_2.47s.mp4", "recording_id": "20200813-subject-02/20200813_154623", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "", "id": "6e86f351667f5a91324257bd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/6e86f351667f5a91324257bd_MANO_GT_2.47s.mp4", "bytes": 789975, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.2687375045879886, "p90": 1.0709693887432135, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.2687375045879886, "p90": 1.0709693887432135, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.2687375045879916, "p90": 1.0709693887432137, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "wms": { "median": 0.9972008050585246, "p90": 0.9990752402013142, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8219178082191781 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 1.0709693887432135, "hand_camera_speed_p90_m_s": 1.0709693887432137, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0709693887432137, "rham_p90_deg_s": null, "wms_p90": 0.9990752402013142, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "b87fe25049b1443c6890656ae2d3c9cb1a7bade93d8fcc47b6fcbc2ef5d561a7", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/6e86f351667f5a91324257bd/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "295b40cd3241fed5af886a0aeaa5fba38398416d1af052e46cf19137084ab6b4", "mano_sha256": "ae2200ba6968fed1ed4872f1038ec4b9cebb14d7ce22d6e3d8cf481f0b2b4958", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "7079ed2860976027ae9e17c1", "sample_id": "7079ed2860976027ae9e17c1", "file": "short_originals_under_20s/DexYCB/7079ed2860976027ae9e17c1_MANO_GT_2.40s.mp4", "recording_id": "20200903-subject-04/20200903_112037", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 11/72 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 15 frames", "id": "7079ed2860976027ae9e17c1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/7079ed2860976027ae9e17c1_MANO_GT_2.40s.mp4", "bytes": 749753, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.4770833622827793, "p90": 1.0462881099057106, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8450704225352113 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8450704225352113 }, "rhcm_continuous_estimate_m_s": { "median": 0.4770833622827793, "p90": 1.0462881099057106, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8450704225352113 }, "rhcm_discrete_m_s": { "median": 0.47708336228277887, "p90": 1.0462881099057084, "valid_pairs": 60, "valid_duration_s": 2.0, "coverage": 0.8450704225352113 }, "wms": { "median": 0.9979306791722949, "p90": 0.9990494956354081, "valid_pairs": 59, "valid_duration_s": 1.9666666666666668, "coverage": 0.8309859154929577 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0462881099057106, "hand_camera_speed_p90_m_s": 1.0462881099057084, "hand_world_valid_fraction": 0.4236111111111111, "hand_camera_valid_fraction": 0.4236111111111111, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0462881099057084, "rham_p90_deg_s": null, "wms_p90": 0.9990494956354081, "gt_coverage": 0.8450704225352113, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "504c8889681c3815fb45aa42a5db9d8b4a39f351c8e889d2d34d795d1d5aee48", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/7079ed2860976027ae9e17c1/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "905d02b59c06ab5f8a6a638477bfa40ed240b8767b4daa096867a41662d028dd", "mano_sha256": "baefb13cf9b023c20242a1235ab55869a0bcc9fa9af5467b836a454ce676c6a8", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "74b4f924c6c02d1f93b396da", "sample_id": "74b4f924c6c02d1f93b396da", "file": "short_originals_under_20s/DexYCB/74b4f924c6c02d1f93b396da_MANO_GT_2.40s.mp4", "recording_id": "20201002-subject-08/20201002_112433", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 21/72 output frames; may include genuine out-of-view frames", "id": "74b4f924c6c02d1f93b396da", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/74b4f924c6c02d1f93b396da_MANO_GT_2.40s.mp4", "bytes": 731448, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.9704597497539403e-13, "p90": 2.97045974975395e-13, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.19703270618696297, "p90": 1.0786480038013564, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "cmh_m_s": { "median": 3.202363323135545e-15, "p90": 4.043935880737184e-15, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "rhcm_continuous_estimate_m_s": { "median": 0.19703270618696067, "p90": 1.0786480038013562, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "rhcm_discrete_m_s": { "median": 0.19703270087257735, "p90": 1.0786480144640225, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "wms": { "median": 0.9949363059303731, "p90": 0.9990737287709663, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 2.97045974975395e-13, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0786480038013564, "hand_camera_speed_p90_m_s": 1.0786480144640225, "hand_world_valid_fraction": 0.3541666666666667, "hand_camera_valid_fraction": 0.3541666666666667, "cmh_p90_m_s": 4.043935880737184e-15, "rhcm_p90_m_s": 1.0786480144640225, "rham_p90_deg_s": null, "wms_p90": 0.9990737287709663, "gt_coverage": 0.7042253521126761, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "83c75568541e4b253318b38fbee951f89c34f346a619f3d38934927bb48af9dc", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/74b4f924c6c02d1f93b396da/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2a8a984bf62d9d377dc92a16443b017e896fd80ad411db74eb45e89f65b680f8", "mano_sha256": "9525d8aefb2b9321c183fdc2cb58ffd2599ee22940222c77098f94d71f417769", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "7f3e8769eb95a6a942159b8d", "sample_id": "7f3e8769eb95a6a942159b8d", "file": "short_originals_under_20s/DexYCB/7f3e8769eb95a6a942159b8d_MANO_GT_2.40s.mp4", "recording_id": "20201022-subject-10/20201022_111832", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 15/72 output frames; may include genuine out-of-view frames", "id": "7f3e8769eb95a6a942159b8d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/7f3e8769eb95a6a942159b8d_MANO_GT_2.40s.mp4", "bytes": 811486, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.3798967309504283, "p90": 1.0605089134800845, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "rhcm_continuous_estimate_m_s": { "median": 0.3798967309504283, "p90": 1.0605089134800845, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "rhcm_discrete_m_s": { "median": 0.3798967309504296, "p90": 1.060508913480084, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "wms": { "median": 0.9973745020300643, "p90": 0.9990579432625688, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0605089134800845, "hand_camera_speed_p90_m_s": 1.060508913480084, "hand_world_valid_fraction": 0.3958333333333333, "hand_camera_valid_fraction": 0.3958333333333333, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.060508913480084, "rham_p90_deg_s": null, "wms_p90": 0.9990579432625688, "gt_coverage": 0.7887323943661971, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "2f7090cd03a81c287c32a5e079ce0a439bc3c222695447abcafb7f2d0cc0cf9d", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/7f3e8769eb95a6a942159b8d/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "863a386eb0a0334226d8d58dc76e647c5d2f8142435962630a867aaa24adbda0", "mano_sha256": "f9081affe735e5ed666b07d33bf4afe75f6c6e1a43036f7097d3c2dfc15c981f", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "7fcaf5c05d9f90959473759a", "sample_id": "7fcaf5c05d9f90959473759a", "file": "short_originals_under_20s/DexYCB/7fcaf5c05d9f90959473759a_MANO_GT_2.40s.mp4", "recording_id": "20201022-subject-10/20201022_113235", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 15/72 output frames; may include genuine out-of-view frames", "id": "7fcaf5c05d9f90959473759a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/7fcaf5c05d9f90959473759a_MANO_GT_2.40s.mp4", "bytes": 732134, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.35600151766972965, "p90": 1.0095973747066815, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "rhcm_continuous_estimate_m_s": { "median": 0.35600151766972965, "p90": 1.0095973747066815, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "rhcm_discrete_m_s": { "median": 0.35600151766972976, "p90": 1.0095973747066833, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "wms": { "median": 0.9971860698589088, "p90": 0.9990094680545349, "valid_pairs": 56, "valid_duration_s": 1.8666666666666667, "coverage": 0.7887323943661971 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0095973747066815, "hand_camera_speed_p90_m_s": 1.0095973747066833, "hand_world_valid_fraction": 0.3958333333333333, "hand_camera_valid_fraction": 0.3958333333333333, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0095973747066833, "rham_p90_deg_s": null, "wms_p90": 0.9990094680545349, "gt_coverage": 0.7887323943661971, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "113009e7ad293d97210a861d3dc073f92f61c20df107fca22be97f57cd637bf9", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/7fcaf5c05d9f90959473759a/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a724e6a8a1f7c46e1461c840dfe4164206be9318d8a83fbdd4ea243659121fd2", "mano_sha256": "25300a2670c648924d459ad324eb9710d70e094a5d45266f86caf31d54b08a42", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "86860155969f8bc95eefc54c", "sample_id": "86860155969f8bc95eefc54c", "file": "short_originals_under_20s/DexYCB/86860155969f8bc95eefc54c_MANO_GT_2.40s.mp4", "recording_id": "20200903-subject-04/20200903_113734", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 23/72 output frames; may include genuine out-of-view frames", "id": "86860155969f8bc95eefc54c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/86860155969f8bc95eefc54c_MANO_GT_2.40s.mp4", "bytes": 796051, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.41443651898376366, "p90": 0.8901638021342628, "valid_pairs": 48, "valid_duration_s": 1.6, "coverage": 0.676056338028169 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 48, "valid_duration_s": 1.6, "coverage": 0.676056338028169 }, "rhcm_continuous_estimate_m_s": { "median": 0.41443651898376366, "p90": 0.8901638021342628, "valid_pairs": 48, "valid_duration_s": 1.6, "coverage": 0.676056338028169 }, "rhcm_discrete_m_s": { "median": 0.4144365189837631, "p90": 0.8901638021342626, "valid_pairs": 48, "valid_duration_s": 1.6, "coverage": 0.676056338028169 }, "wms": { "median": 0.997630253077964, "p90": 0.998885268955363, "valid_pairs": 47, "valid_duration_s": 1.5666666666666664, "coverage": 0.6619718309859154 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.8901638021342628, "hand_camera_speed_p90_m_s": 0.8901638021342626, "hand_world_valid_fraction": 0.3402777777777778, "hand_camera_valid_fraction": 0.3402777777777778, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.8901638021342626, "rham_p90_deg_s": null, "wms_p90": 0.998885268955363, "gt_coverage": 0.676056338028169, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "6a75e0265935298498bbf038029f822cafff109fd9bcf4e0e894804d0b4e6a38", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/86860155969f8bc95eefc54c/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8d899eabcd6efb96c57c55426a2e131964f1ac18750e6993878e3273e9ec2c4a", "mano_sha256": "d046df1da1e45154a7a0c6325a98f4c08318f21241e56be6760873e392cb161e", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "8e57aabd56f536c1b0e6e46c", "sample_id": "8e57aabd56f536c1b0e6e46c", "file": "short_originals_under_20s/DexYCB/8e57aabd56f536c1b0e6e46c_MANO_GT_2.40s.mp4", "recording_id": "20200709-subject-01/20200709_144814", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "", "id": "8e57aabd56f536c1b0e6e46c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/8e57aabd56f536c1b0e6e46c_MANO_GT_2.40s.mp4", "bytes": 774350, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.20433371362178476, "p90": 0.8239342057533315, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.20433371362178476, "p90": 0.8239342057533315, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.2043337136217847, "p90": 0.8239342057533329, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "wms": { "median": 0.9951298791495976, "p90": 0.9987877821127725, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.8239342057533315, "hand_camera_speed_p90_m_s": 0.8239342057533329, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.8239342057533329, "rham_p90_deg_s": null, "wms_p90": 0.9987877821127725, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "92abc6f3ef7691cc1f558033e991c21018e66670be8ed830d584cec2dd966be5", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/8e57aabd56f536c1b0e6e46c/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ab66dace790415df896e6a655a21f2d968038a69f1e76b3b9799ea75c62023f0", "mano_sha256": "a97b9d0220eecb47d823deb0278f9296abcbeea4496d0a4a9d7b1a3a6a08510d", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "9f3b17cd576b35ace407aa70", "sample_id": "9f3b17cd576b35ace407aa70", "file": "short_originals_under_20s/DexYCB/9f3b17cd576b35ace407aa70_MANO_GT_2.47s.mp4", "recording_id": "20200820-subject-03/20200820_135648", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 1/74 output frames; may include genuine out-of-view frames", "id": "9f3b17cd576b35ace407aa70", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/9f3b17cd576b35ace407aa70_MANO_GT_2.47s.mp4", "bytes": 886513, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.1020026809811092, "p90": 0.9381258574446147, "valid_pairs": 72, "valid_duration_s": 2.3999999999999995, "coverage": 0.9863013698630135 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 72, "valid_duration_s": 2.3999999999999995, "coverage": 0.9863013698630135 }, "rhcm_continuous_estimate_m_s": { "median": 0.1020026809811092, "p90": 0.9381258574446147, "valid_pairs": 72, "valid_duration_s": 2.3999999999999995, "coverage": 0.9863013698630135 }, "rhcm_discrete_m_s": { "median": 0.1020026809811087, "p90": 0.9381258574446146, "valid_pairs": 72, "valid_duration_s": 2.3999999999999995, "coverage": 0.9863013698630135 }, "wms": { "median": 0.9956462375874514, "p90": 0.999135251427482, "valid_pairs": 49, "valid_duration_s": 1.633333333333333, "coverage": 0.6712328767123287 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.9381258574446147, "hand_camera_speed_p90_m_s": 0.9381258574446146, "hand_world_valid_fraction": 0.49324324324324326, "hand_camera_valid_fraction": 0.49324324324324326, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.9381258574446146, "rham_p90_deg_s": null, "wms_p90": 0.999135251427482, "gt_coverage": 0.9863013698630135, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "15b8401f723396b020e114804dc6da65101bae9ad2466e428274739338ade99c", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/9f3b17cd576b35ace407aa70/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5364d279e0dc111f284ce572a72b28eaf3dc9a1abb61314a9296225f4fc591e8", "mano_sha256": "e211f0423f5116f1a0e732365fe8002b98dec63c21a65dd87140400470968eb8", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "a1d48a65b042cd3c582ad1fd", "sample_id": "a1d48a65b042cd3c582ad1fd", "file": "short_originals_under_20s/DexYCB/a1d48a65b042cd3c582ad1fd_MANO_GT_2.40s.mp4", "recording_id": "20201022-subject-10/20201022_114824", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 25/72 output frames; may include genuine out-of-view frames", "id": "a1d48a65b042cd3c582ad1fd", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/a1d48a65b042cd3c582ad1fd_MANO_GT_2.40s.mp4", "bytes": 645369, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.36957586582978397, "p90": 1.2043142890237872, "valid_pairs": 46, "valid_duration_s": 1.5333333333333332, "coverage": 0.6478873239436619 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 46, "valid_duration_s": 1.5333333333333332, "coverage": 0.6478873239436619 }, "rhcm_continuous_estimate_m_s": { "median": 0.36957586582978397, "p90": 1.2043142890237872, "valid_pairs": 46, "valid_duration_s": 1.5333333333333332, "coverage": 0.6478873239436619 }, "rhcm_discrete_m_s": { "median": 0.36957586582978397, "p90": 1.2043142890237872, "valid_pairs": 46, "valid_duration_s": 1.5333333333333332, "coverage": 0.6478873239436619 }, "wms": { "median": 0.9973014577320738, "p90": 0.9991701802212162, "valid_pairs": 46, "valid_duration_s": 1.5333333333333332, "coverage": 0.6478873239436619 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.2043142890237872, "hand_camera_speed_p90_m_s": 1.2043142890237872, "hand_world_valid_fraction": 0.3263888888888889, "hand_camera_valid_fraction": 0.3263888888888889, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.2043142890237872, "rham_p90_deg_s": null, "wms_p90": 0.9991701802212162, "gt_coverage": 0.6478873239436619, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "5c03f17279cae24ba61379e2ebeb9ba6c803caf79f4a8bc1910b1fbbf9b9711d", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/a1d48a65b042cd3c582ad1fd/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "6f3f075a8b6af94421b572eff9cf936cb348607372eb971a324a79c6a4907e96", "mano_sha256": "20ebe90e1e4234b801cc5a086ac9f4e28c67a500fd2f2ca9eead570b686d6751", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "ac6a877d9c76c9abc3cdb63d", "sample_id": "ac6a877d9c76c9abc3cdb63d", "file": "short_originals_under_20s/DexYCB/ac6a877d9c76c9abc3cdb63d_MANO_GT_2.47s.mp4", "recording_id": "20200928-subject-07/20200928_154423", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 21/74 output frames; may include genuine out-of-view frames", "id": "ac6a877d9c76c9abc3cdb63d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/ac6a877d9c76c9abc3cdb63d_MANO_GT_2.47s.mp4", "bytes": 733216, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.11222563742304979, "p90": 0.9150329985042446, "valid_pairs": 52, "valid_duration_s": 1.7333333333333332, "coverage": 0.7123287671232876 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 52, "valid_duration_s": 1.7333333333333332, "coverage": 0.7123287671232876 }, "rhcm_continuous_estimate_m_s": { "median": 0.11222563742304979, "p90": 0.9150329985042446, "valid_pairs": 52, "valid_duration_s": 1.7333333333333332, "coverage": 0.7123287671232876 }, "rhcm_discrete_m_s": { "median": 0.11222563742304939, "p90": 0.9150329985042437, "valid_pairs": 52, "valid_duration_s": 1.7333333333333332, "coverage": 0.7123287671232876 }, "wms": { "median": 0.9929768710970543, "p90": 0.998973583251997, "valid_pairs": 46, "valid_duration_s": 1.5333333333333334, "coverage": 0.6301369863013699 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.9150329985042446, "hand_camera_speed_p90_m_s": 0.9150329985042437, "hand_world_valid_fraction": 0.3581081081081081, "hand_camera_valid_fraction": 0.3581081081081081, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.9150329985042437, "rham_p90_deg_s": null, "wms_p90": 0.998973583251997, "gt_coverage": 0.7123287671232876, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "cc4c13f5e3444e36520bfbd74a04eb6ba8e6334ccb22fe691bbda0ec7cc3b6b5", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/ac6a877d9c76c9abc3cdb63d/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "28836274c8c6d70ceb5f1f47337c68345768a5b0b26c98fd1f2edfb754c8bda5", "mano_sha256": "c1b65e58da95bc9e7a3d13eafaf8be1282145540353abb839ab99c7d81d3ab77", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "aeb2f4db7390bc5774d912d8", "sample_id": "aeb2f4db7390bc5774d912d8", "file": "short_originals_under_20s/DexYCB/aeb2f4db7390bc5774d912d8_MANO_GT_2.47s.mp4", "recording_id": "20201002-subject-08/20201002_110425", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_test", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 19/74 output frames; may include genuine out-of-view frames", "id": "aeb2f4db7390bc5774d912d8", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/aeb2f4db7390bc5774d912d8_MANO_GT_2.47s.mp4", "bytes": 866963, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.22229189397136095, "p90": 0.9118196258023441, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "rhcm_continuous_estimate_m_s": { "median": 0.22229189397136095, "p90": 0.9118196258023441, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "rhcm_discrete_m_s": { "median": 0.22229189397136093, "p90": 0.911819625802344, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "wms": { "median": 0.9956249849665237, "p90": 0.9989154679961049, "valid_pairs": 50, "valid_duration_s": 1.6666666666666665, "coverage": 0.684931506849315 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.9118196258023441, "hand_camera_speed_p90_m_s": 0.911819625802344, "hand_world_valid_fraction": 0.3716216216216216, "hand_camera_valid_fraction": 0.3716216216216216, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.911819625802344, "rham_p90_deg_s": null, "wms_p90": 0.9989154679961049, "gt_coverage": 0.7397260273972602, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "4b9babd331626ec956d40d476d56297c8a2317cb74d823f148963825ccee9780", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/aeb2f4db7390bc5774d912d8/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "2a73fd3a1cc25a40c41cc93b10ebd37956959b34055b02d999b7c123f8d08423", "mano_sha256": "ad4626256fb7b6043c154c87e1835a2c9e8c1e5ca81969c9b53307ba395bef61", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "b2688f3c538c45a7b8acc203", "sample_id": "b2688f3c538c45a7b8acc203", "file": "short_originals_under_20s/DexYCB/b2688f3c538c45a7b8acc203_MANO_GT_2.40s.mp4", "recording_id": "20201002-subject-08/20201002_112718", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 3/72 output frames; may include genuine out-of-view frames", "id": "b2688f3c538c45a7b8acc203", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/b2688f3c538c45a7b8acc203_MANO_GT_2.40s.mp4", "bytes": 796941, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.29703205021174095, "p90": 1.0071104864142453, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "rhcm_continuous_estimate_m_s": { "median": 0.29703205021174095, "p90": 1.0071104864142453, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "rhcm_discrete_m_s": { "median": 0.2970320502117408, "p90": 1.007110486414244, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "wms": { "median": 0.9967095980915907, "p90": 0.999020957926344, "valid_pairs": 65, "valid_duration_s": 2.166666666666667, "coverage": 0.915492957746479 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0071104864142453, "hand_camera_speed_p90_m_s": 1.007110486414244, "hand_world_valid_fraction": 0.4791666666666667, "hand_camera_valid_fraction": 0.4791666666666667, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.007110486414244, "rham_p90_deg_s": null, "wms_p90": 0.999020957926344, "gt_coverage": 0.9577464788732394, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "cec03155cc83e3f548934d311c9f24bf9faf7763735afd1ccf9446a2ea16698f", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/b2688f3c538c45a7b8acc203/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "bc1eb41e6a79fa7ea0074deb280313b21e0365bc2828e013b8a076c6d4c1205d", "mano_sha256": "fe678610232733c28f057182307010926ac498d6b006f71bc12e15306c05cfa9", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "b710ad273d403cf9622886ca", "sample_id": "b710ad273d403cf9622886ca", "file": "short_originals_under_20s/DexYCB/b710ad273d403cf9622886ca_MANO_GT_2.40s.mp4", "recording_id": "20201015-subject-09/20201015_144823", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "", "id": "b710ad273d403cf9622886ca", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/b710ad273d403cf9622886ca_MANO_GT_2.40s.mp4", "bytes": 711614, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.08477513986803324, "p90": 0.7293941048602082, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.08477513986803324, "p90": 0.7293941048602082, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.08477513986803216, "p90": 0.7293941048602088, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "wms": { "median": 0.9938447883993626, "p90": 0.9986490513779607, "valid_pairs": 62, "valid_duration_s": 2.066666666666667, "coverage": 0.8732394366197184 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.7293941048602082, "hand_camera_speed_p90_m_s": 0.7293941048602088, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.7293941048602088, "rham_p90_deg_s": null, "wms_p90": 0.9986490513779607, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "19d3d939ec983fb0ccf20edbbce34f5560a7b67c10df4c46d16b9a4f29f4ac0a", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/b710ad273d403cf9622886ca/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "288b6a5075a7b755600ecd82ad59c79fec3e87a133a2ddb3ff59ccdf12613535", "mano_sha256": "2121e6c92de0d0b7b16bc178350c10590302a96321c3e75098b8517325733691", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "b7763e84a0c660b613f92d7c", "sample_id": "b7763e84a0c660b613f92d7c", "file": "short_originals_under_20s/DexYCB/b7763e84a0c660b613f92d7c_MANO_GT_2.40s.mp4", "recording_id": "20200813-subject-02/20200813_154659", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "", "id": "b7763e84a0c660b613f92d7c", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/b7763e84a0c660b613f92d7c_MANO_GT_2.40s.mp4", "bytes": 913465, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.14953261941659923, "p90": 0.666165047603323, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.14953261941659923, "p90": 0.666165047603323, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.14953261941659604, "p90": 0.6661650476033231, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "wms": { "median": 0.9966290251779812, "p90": 0.9988159064486276, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.666165047603323, "hand_camera_speed_p90_m_s": 0.6661650476033231, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.6661650476033231, "rham_p90_deg_s": null, "wms_p90": 0.9988159064486276, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "bf12d0faea5aa40d4629a575f06a9c3a758c1dec588505ee9374d28c08567db4", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/b7763e84a0c660b613f92d7c/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "4e0e242ced77f47bf1b2d649096e5878cc0c20490b0e0cf1852dd4bfbd85471e", "mano_sha256": "533f21967757993aad7e089a9f91b0dfe8a32ff368048fd37e6de95177b87070", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "b8d7f5abaed597f996acba59", "sample_id": "b8d7f5abaed597f996acba59", "file": "short_originals_under_20s/DexYCB/b8d7f5abaed597f996acba59_MANO_GT_2.47s.mp4", "recording_id": "20200903-subject-04/20200903_113804", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 25/74 output frames; may include genuine out-of-view frames", "id": "b8d7f5abaed597f996acba59", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/b8d7f5abaed597f996acba59_MANO_GT_2.47s.mp4", "bytes": 869329, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 1.0578268581810926e-13, "p90": 1.0578268581810963e-13, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.5568139822055951, "p90": 1.0575387113283892, "valid_pairs": 48, "valid_duration_s": 1.5999999999999996, "coverage": 0.6575342465753424 }, "cmh_m_s": { "median": 1.303308620531288e-15, "p90": 1.7009532693654272e-15, "valid_pairs": 48, "valid_duration_s": 1.5999999999999996, "coverage": 0.6575342465753424 }, "rhcm_continuous_estimate_m_s": { "median": 0.5568139822055957, "p90": 1.0575387113283896, "valid_pairs": 48, "valid_duration_s": 1.5999999999999996, "coverage": 0.6575342465753424 }, "rhcm_discrete_m_s": { "median": 0.5568139451979295, "p90": 1.0575385670111201, "valid_pairs": 48, "valid_duration_s": 1.5999999999999996, "coverage": 0.6575342465753424 }, "wms": { "median": 0.9982071354316544, "p90": 0.9990546674168357, "valid_pairs": 48, "valid_duration_s": 1.5999999999999996, "coverage": 0.6575342465753424 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 1.0578268581810963e-13, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 1.0575387113283892, "hand_camera_speed_p90_m_s": 1.0575385670111201, "hand_world_valid_fraction": 0.3310810810810811, "hand_camera_valid_fraction": 0.3310810810810811, "cmh_p90_m_s": 1.7009532693654272e-15, "rhcm_p90_m_s": 1.0575385670111201, "rham_p90_deg_s": null, "wms_p90": 0.9990546674168357, "gt_coverage": 0.6575342465753424, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "ce2cbd9244861f219e9d7ed148a3b4745c945abf45406e7b8c11e2bf035fca22", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/b8d7f5abaed597f996acba59/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "0a19777e784310af8023cc31b11f32375769e683c00448962bc9c947569db7d6", "mano_sha256": "878919ff5861ddc342a42c89c0466ca2f558b76643dcd8831a29ad287cb19a4c", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "b90ffb921e005cf4dbf20075", "sample_id": "b90ffb921e005cf4dbf20075", "file": "short_originals_under_20s/DexYCB/b90ffb921e005cf4dbf20075_MANO_GT_2.40s.mp4", "recording_id": "20200908-subject-05/20200908_152530", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_test", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 7/72 output frames; may include genuine out-of-view frames", "id": "b90ffb921e005cf4dbf20075", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/b90ffb921e005cf4dbf20075_MANO_GT_2.40s.mp4", "bytes": 836819, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 1.2985338845361775e-14, "p90": 1.2985338845361818e-14, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.25146694466607095, "p90": 1.0601634390460752, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "cmh_m_s": { "median": 1.452358943841851e-16, "p90": 2.062897260025142e-16, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rhcm_continuous_estimate_m_s": { "median": 0.2514669446660709, "p90": 1.060163439046075, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rhcm_discrete_m_s": { "median": 0.2514669370949546, "p90": 1.06016336772192, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "wms": { "median": 0.996038896762969, "p90": 0.9990576276430857, "valid_pairs": 64, "valid_duration_s": 2.1333333333333333, "coverage": 0.9014084507042254 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 1.2985338845361818e-14, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.0601634390460752, "hand_camera_speed_p90_m_s": 1.06016336772192, "hand_world_valid_fraction": 0.4513888888888889, "hand_camera_valid_fraction": 0.4513888888888889, "cmh_p90_m_s": 2.062897260025142e-16, "rhcm_p90_m_s": 1.06016336772192, "rham_p90_deg_s": null, "wms_p90": 0.9990576276430857, "gt_coverage": 0.9014084507042254, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "ae0b3a1aa2b802c3bb193148bddde416f7650b21430023a9801ba4a7b10ecdb0", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/b90ffb921e005cf4dbf20075/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "9b38ff3a1ea8bb7c85c7090ded8daf5653f5d5b5d47285abd948a64747d83952", "mano_sha256": "874a927403178cf10f0f051a6a76c7d2073f0e18e8405d8a8f688b79f805e99c", "duration_s": 2.334, "frames": 70, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "bb5c6afee18a4dad7aa8da96", "sample_id": "bb5c6afee18a4dad7aa8da96", "file": "short_originals_under_20s/DexYCB/bb5c6afee18a4dad7aa8da96_MANO_GT_2.33s.mp4", "recording_id": "20201002-subject-08/20201002_112902", "source_first_frame": "0", "source_last_frame": "69", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 9/70 output frames; may include genuine out-of-view frames", "id": "bb5c6afee18a4dad7aa8da96", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/bb5c6afee18a4dad7aa8da96_MANO_GT_2.33s.mp4", "bytes": 765037, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 69, "valid_duration_s": 2.3, "coverage": 1.0 }, "camera_rotation": { "median": 2.9704597497539403e-13, "p90": 2.97045974975395e-13, "valid_pairs": 69, "valid_duration_s": 2.3, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.20727914552491328, "p90": 1.2103500486583583, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.8695652173913043 }, "cmh_m_s": { "median": 2.8075744409774972e-15, "p90": 4.158920092085575e-15, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.8695652173913043 }, "rhcm_continuous_estimate_m_s": { "median": 0.2072791455249115, "p90": 1.210350048658359, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.8695652173913043 }, "rhcm_discrete_m_s": { "median": 0.20727914408292947, "p90": 1.210350071856128, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.8695652173913043 }, "wms": { "median": 0.9951987063424755, "p90": 0.9991743569960891, "valid_pairs": 60, "valid_duration_s": 1.9999999999999998, "coverage": 0.8695652173913043 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 2.97045974975395e-13, "camera_valid_fraction": 1.0, "camera_pairs": 69, "hand_world_speed_p90_m_s": 1.2103500486583583, "hand_camera_speed_p90_m_s": 1.210350071856128, "hand_world_valid_fraction": 0.4357142857142857, "hand_camera_valid_fraction": 0.4357142857142857, "cmh_p90_m_s": 4.158920092085575e-15, "rhcm_p90_m_s": 1.210350071856128, "rham_p90_deg_s": null, "wms_p90": 0.9991743569960891, "gt_coverage": 0.8695652173913043, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "2f486b0b3bbe4781598947211f08c11e32efe1950b79fd5ba457bfeb420e6c15", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/bb5c6afee18a4dad7aa8da96/reference.npz", "window_s": [ 0.0, 2.3 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 70, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "a1d1696b4aa7987f2326b49fa1c294c8e9296ba52eb72f70de2a5069177895d0", "mano_sha256": "68de2323b5a9b0aa91f7d83329a906b2f73ac974caf18352bf39380802da784f", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "bbbff1bdce22da08321b6a33", "sample_id": "bbbff1bdce22da08321b6a33", "file": "short_originals_under_20s/DexYCB/bbbff1bdce22da08321b6a33_MANO_GT_2.40s.mp4", "recording_id": "20200908-subject-05/20200908_144505", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 1/72 output frames; may include genuine out-of-view frames", "id": "bbbff1bdce22da08321b6a33", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/bbbff1bdce22da08321b6a33_MANO_GT_2.40s.mp4", "bytes": 825553, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.16172678346442648, "p90": 0.6436788592511409, "valid_pairs": 70, "valid_duration_s": 2.333333333333333, "coverage": 0.9859154929577464 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 70, "valid_duration_s": 2.333333333333333, "coverage": 0.9859154929577464 }, "rhcm_continuous_estimate_m_s": { "median": 0.16172678346442648, "p90": 0.6436788592511409, "valid_pairs": 70, "valid_duration_s": 2.333333333333333, "coverage": 0.9859154929577464 }, "rhcm_discrete_m_s": { "median": 0.16172678346442593, "p90": 0.6436788592511395, "valid_pairs": 70, "valid_duration_s": 2.333333333333333, "coverage": 0.9859154929577464 }, "wms": { "median": 0.9950288176271479, "p90": 0.9984795562950335, "valid_pairs": 63, "valid_duration_s": 2.0999999999999996, "coverage": 0.8873239436619716 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.6436788592511409, "hand_camera_speed_p90_m_s": 0.6436788592511395, "hand_world_valid_fraction": 0.4930555555555556, "hand_camera_valid_fraction": 0.4930555555555556, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.6436788592511395, "rham_p90_deg_s": null, "wms_p90": 0.9984795562950335, "gt_coverage": 0.9859154929577464, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "8ad96a9694e2d28995dc26f01d31bcbbb784f4fe4242955508bb94e1f83b29a7", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/bbbff1bdce22da08321b6a33/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "5a5b683fd369c6289739ee24b80b1d55ff9fc300aa9b8dd4a0d3146ac9870e69", "mano_sha256": "918ef6a2d6f8241dc452c58888de801b10d3ca4f6c7eba3a673194d7e3203837", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "bbc37b3adf677609ac1c0f64", "sample_id": "bbc37b3adf677609ac1c0f64", "file": "short_originals_under_20s/DexYCB/bbc37b3adf677609ac1c0f64_MANO_GT_2.47s.mp4", "recording_id": "20200820-subject-03/20200820_141302", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_test", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 3/74 output frames; may include genuine out-of-view frames", "id": "bbc37b3adf677609ac1c0f64", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/bbc37b3adf677609ac1c0f64_MANO_GT_2.47s.mp4", "bytes": 770719, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.17206700098247224, "p90": 1.0793872481291809, "valid_pairs": 70, "valid_duration_s": 2.333333333333333, "coverage": 0.958904109589041 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 70, "valid_duration_s": 2.333333333333333, "coverage": 0.958904109589041 }, "rhcm_continuous_estimate_m_s": { "median": 0.17206700098247224, "p90": 1.0793872481291809, "valid_pairs": 70, "valid_duration_s": 2.333333333333333, "coverage": 0.958904109589041 }, "rhcm_discrete_m_s": { "median": 0.1720670009824719, "p90": 1.0793872481291804, "valid_pairs": 70, "valid_duration_s": 2.333333333333333, "coverage": 0.958904109589041 }, "wms": { "median": 0.9974668926589894, "p90": 0.9991151001784059, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.6849315068493151 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 1.0793872481291809, "hand_camera_speed_p90_m_s": 1.0793872481291804, "hand_world_valid_fraction": 0.4797297297297297, "hand_camera_valid_fraction": 0.4797297297297297, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0793872481291804, "rham_p90_deg_s": null, "wms_p90": 0.9991151001784059, "gt_coverage": 0.958904109589041, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "a07dc49b54039f37c1eda6271e95091acbaaed3a0fa32445517ecdcad86edb74", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/bbc37b3adf677609ac1c0f64/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7724ae4662b6493b42cf12377ec0346a79174c2a8bb0e62f9d4bd6d574fa4c75", "mano_sha256": "8d6edcc9a7c2a756877bb15dd0d48d770b4af2bb780516e4f830dd9108276819", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "c30eceb2ceeb0a7a8915bdc6", "sample_id": "c30eceb2ceeb0a7a8915bdc6", "file": "short_originals_under_20s/DexYCB/c30eceb2ceeb0a7a8915bdc6_MANO_GT_2.40s.mp4", "recording_id": "20200813-subject-02/20200813_150823", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_validation", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "", "id": "c30eceb2ceeb0a7a8915bdc6", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/c30eceb2ceeb0a7a8915bdc6_MANO_GT_2.40s.mp4", "bytes": 820233, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 7.246734943301247e-13, "p90": 7.246734943301271e-13, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.4338906689716083, "p90": 0.9940069261847899, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 5.664636910464709e-15, "p90": 9.059151052220867e-15, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.4338906689716134, "p90": 0.9940069261847906, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.43389064957011486, "p90": 0.994006838619227, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "wms": { "median": 0.9978449876237377, "p90": 0.9989963102802835, "valid_pairs": 67, "valid_duration_s": 2.233333333333333, "coverage": 0.9436619718309858 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 7.246734943301271e-13, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.9940069261847899, "hand_camera_speed_p90_m_s": 0.994006838619227, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 9.059151052220867e-15, "rhcm_p90_m_s": 0.994006838619227, "rham_p90_deg_s": null, "wms_p90": 0.9989963102802835, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "df0f545be160faea0df81b1c4cf3028f3bb35eb131dd3906d33ccc4be2172e0c", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/c30eceb2ceeb0a7a8915bdc6/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "b6c1c9db68b74d9ecc925de29f4bd4d67b92734ea92bc5992836f5fcec984ca9", "mano_sha256": "37d73ac619b2df0e6fef30638cdc648c3d32f44a2be3392889296e8422a579fa", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "c4f3c24273cb8fd5d40b6d42", "sample_id": "c4f3c24273cb8fd5d40b6d42", "file": "short_originals_under_20s/DexYCB/c4f3c24273cb8fd5d40b6d42_MANO_GT_2.47s.mp4", "recording_id": "20200918-subject-06/20200918_114548", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 11/74 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 3 frames", "id": "c4f3c24273cb8fd5d40b6d42", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/c4f3c24273cb8fd5d40b6d42_MANO_GT_2.47s.mp4", "bytes": 820838, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.33492886241078057, "p90": 1.2119463071946925, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8493150684931506 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8493150684931506 }, "rhcm_continuous_estimate_m_s": { "median": 0.33492886241078057, "p90": 1.2119463071946925, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8493150684931506 }, "rhcm_discrete_m_s": { "median": 0.33492886241078124, "p90": 1.2119463071946925, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8493150684931506 }, "wms": { "median": 0.9970873676211265, "p90": 0.9991771002032142, "valid_pairs": 58, "valid_duration_s": 1.9333333333333336, "coverage": 0.7945205479452057 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 1.2119463071946925, "hand_camera_speed_p90_m_s": 1.2119463071946925, "hand_world_valid_fraction": 0.42567567567567566, "hand_camera_valid_fraction": 0.42567567567567566, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.2119463071946925, "rham_p90_deg_s": null, "wms_p90": 0.9991771002032142, "gt_coverage": 0.8493150684931506, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "ebd591942c28da9e862a2141c6d00dc4dc355ecd5f098392369421977b822c87", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/c4f3c24273cb8fd5d40b6d42/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "76eb593e2b841399e54445f903500352df603cb7959d054f5ee6e5734c5e3814", "mano_sha256": "eccb63c779eb4fa7dc7a6265dc7cb362e77f99e9cd857c05baa927de79a6c283", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "外部视角官方 MANO 可见，持物接触近似；这是实际短原片。", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "c7f2986817f450d9790750c5", "sample_id": "c7f2986817f450d9790750c5", "file": "short_originals_under_20s/DexYCB/c7f2986817f450d9790750c5_MANO_GT_2.40s.mp4", "recording_id": "20201015-subject-09/20201015_150342", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "", "id": "c7f2986817f450d9790750c5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/c7f2986817f450d9790750c5_MANO_GT_2.40s.mp4", "bytes": 780190, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 2.964482599567262e-13, "p90": 2.964482599567272e-13, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.20917501917498296, "p90": 0.6190469441994351, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "cmh_m_s": { "median": 9.730626754991448e-16, "p90": 1.281573903496035e-15, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_continuous_estimate_m_s": { "median": 0.20917501917498257, "p90": 0.6190469441994357, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rhcm_discrete_m_s": { "median": 0.2091750266483599, "p90": 0.6190470258195865, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "wms": { "median": 0.9952420606220151, "p90": 0.9983872188882549, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 2.964482599567272e-13, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.6190469441994351, "hand_camera_speed_p90_m_s": 0.6190470258195865, "hand_world_valid_fraction": 0.5, "hand_camera_valid_fraction": 0.5, "cmh_p90_m_s": 1.281573903496035e-15, "rhcm_p90_m_s": 0.6190470258195865, "rham_p90_deg_s": null, "wms_p90": 0.9983872188882549, "gt_coverage": 1.0, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "e0db16ee966af43abc3b3ecb0b0e8f8dba5eab01fc331cf6e18f7760872775de", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/c7f2986817f450d9790750c5/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ba43374ab9d6860087ca35e74c43832254feb60d00f23fb283d00993a0ea43fd", "mano_sha256": "07722daf64c9e424300fb1db91c108af289d26ae87f37bbac8c7ca71a65e1ef0", "duration_s": 2.334, "frames": 70, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "c94b78d91bef98fc41d6432e", "sample_id": "c94b78d91bef98fc41d6432e", "file": "short_originals_under_20s/DexYCB/c94b78d91bef98fc41d6432e_MANO_GT_2.33s.mp4", "recording_id": "20201002-subject-08/20201002_111735", "source_first_frame": "0", "source_last_frame": "69", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 15/70 output frames; may include genuine out-of-view frames", "id": "c94b78d91bef98fc41d6432e", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/c94b78d91bef98fc41d6432e_MANO_GT_2.33s.mp4", "bytes": 795634, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 69, "valid_duration_s": 2.3, "coverage": 1.0 }, "camera_rotation": { "median": 9.87791534864735e-13, "p90": 9.877915348647385e-13, "valid_pairs": 69, "valid_duration_s": 2.3, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.18791008690106517, "p90": 0.9636323628231269, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.782608695652174 }, "cmh_m_s": { "median": 1.7243196611578894e-14, "p90": 1.9585206135242206e-14, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.782608695652174 }, "rhcm_continuous_estimate_m_s": { "median": 0.18791008690107408, "p90": 0.963632362823139, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.782608695652174 }, "rhcm_discrete_m_s": { "median": 0.1879100726176876, "p90": 0.9636322809729474, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.782608695652174 }, "wms": { "median": 0.9947064633443863, "p90": 0.9989630209105912, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.782608695652174 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 9.877915348647385e-13, "camera_valid_fraction": 1.0, "camera_pairs": 69, "hand_world_speed_p90_m_s": 0.9636323628231269, "hand_camera_speed_p90_m_s": 0.9636322809729474, "hand_world_valid_fraction": 0.39285714285714285, "hand_camera_valid_fraction": 0.39285714285714285, "cmh_p90_m_s": 1.9585206135242206e-14, "rhcm_p90_m_s": 0.9636322809729474, "rham_p90_deg_s": null, "wms_p90": 0.9989630209105912, "gt_coverage": 0.782608695652174, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "cdcecd599995ec6562a6e121dfb97a3be56a6bdb2a85486f29fd0880f3a0c2a0", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/c94b78d91bef98fc41d6432e/reference.npz", "window_s": [ 0.0, 2.3 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 70, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ced722de0accafc8df769fa5b82a53cc46533cb86de92dee477d543553742c21", "mano_sha256": "d790ef0254c14f4374d1b376434dbbdb91be49c85204765d6b3f9b974b3c33ae", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "cc91f5e00a5a7941e3e941de", "sample_id": "cc91f5e00a5a7941e3e941de", "file": "short_originals_under_20s/DexYCB/cc91f5e00a5a7941e3e941de_MANO_GT_2.40s.mp4", "recording_id": "20200928-subject-07/20200928_155616", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_test", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 21/72 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 14 frames", "id": "cc91f5e00a5a7941e3e941de", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/cc91f5e00a5a7941e3e941de_MANO_GT_2.40s.mp4", "bytes": 788673, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.17474345546603218, "p90": 1.106965824451809, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "rhcm_continuous_estimate_m_s": { "median": 0.17474345546603218, "p90": 1.106965824451809, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "rhcm_discrete_m_s": { "median": 0.17474345546603154, "p90": 1.106965824451811, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "wms": { "median": 0.9943067133089143, "p90": 0.9990974363242785, "valid_pairs": 50, "valid_duration_s": 1.6666666666666667, "coverage": 0.7042253521126761 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 1.106965824451809, "hand_camera_speed_p90_m_s": 1.106965824451811, "hand_world_valid_fraction": 0.3541666666666667, "hand_camera_valid_fraction": 0.3541666666666667, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.106965824451811, "rham_p90_deg_s": null, "wms_p90": 0.9990974363242785, "gt_coverage": 0.7042253521126761, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "972ef862946a6be604dcbcf1e1bc2d2582500a3a56f47b37b81ee267aab079e9", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/cc91f5e00a5a7941e3e941de/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "ed048c3831f5a52b9984744c247b44b8bfa73c50ae7ef59582c762ce1e5af0d1", "mano_sha256": "72d3b1c470b6ff115336a163c51a8b845a1d719ceb2f3b479ef71c71c956326d", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "d472290630facdd8af436646", "sample_id": "d472290630facdd8af436646", "file": "short_originals_under_20s/DexYCB/d472290630facdd8af436646_MANO_GT_2.47s.mp4", "recording_id": "20200918-subject-06/20200918_115727", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 9/74 output frames; may include genuine out-of-view frames", "id": "d472290630facdd8af436646", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/d472290630facdd8af436646_MANO_GT_2.47s.mp4", "bytes": 798585, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.2300805071727211, "p90": 1.3754563609156587, "valid_pairs": 64, "valid_duration_s": 2.133333333333333, "coverage": 0.8767123287671231 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 64, "valid_duration_s": 2.133333333333333, "coverage": 0.8767123287671231 }, "rhcm_continuous_estimate_m_s": { "median": 0.2300805071727211, "p90": 1.3754563609156587, "valid_pairs": 64, "valid_duration_s": 2.133333333333333, "coverage": 0.8767123287671231 }, "rhcm_discrete_m_s": { "median": 0.23008050717272133, "p90": 1.375456360915658, "valid_pairs": 64, "valid_duration_s": 2.133333333333333, "coverage": 0.8767123287671231 }, "wms": { "median": 0.9956718907703763, "p90": 0.9992734846888918, "valid_pairs": 64, "valid_duration_s": 2.133333333333333, "coverage": 0.8767123287671231 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 1.3754563609156587, "hand_camera_speed_p90_m_s": 1.375456360915658, "hand_world_valid_fraction": 0.4391891891891892, "hand_camera_valid_fraction": 0.4391891891891892, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.375456360915658, "rham_p90_deg_s": null, "wms_p90": 0.9992734846888918, "gt_coverage": 0.8767123287671231, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "c150b94b2db7d530907bac47eb40db86ce0c2b16778d7c8029fa66e5f245a99f", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/d472290630facdd8af436646/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "070307fdc903a9bb555793090a08b58df824d18a6c40c1fb4840fd845cc057da", "mano_sha256": "29fe8f62683b678ae981b06659c862702ad4d08a8f4ba5a3b805cf3b0f7fa90e", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "evaluation_only", "context_id": "d7c705b7c5512fb693d4709d", "sample_id": "d7c705b7c5512fb693d4709d", "file": "short_originals_under_20s/DexYCB/d7c705b7c5512fb693d4709d_MANO_GT_2.40s.mp4", "recording_id": "20200820-subject-03/20200820_144800", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_test", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 9/72 output frames; may include genuine out-of-view frames", "id": "d7c705b7c5512fb693d4709d", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/d7c705b7c5512fb693d4709d_MANO_GT_2.40s.mp4", "bytes": 783495, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.22339663129850784, "p90": 0.8557899467218846, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "rhcm_continuous_estimate_m_s": { "median": 0.22339663129850784, "p90": 0.8557899467218846, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "rhcm_discrete_m_s": { "median": 0.22339663129850768, "p90": 0.8557899467218826, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "wms": { "median": 0.9967078508673808, "p90": 0.998881262098843, "valid_pairs": 54, "valid_duration_s": 1.7999999999999996, "coverage": 0.76056338028169 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.8557899467218846, "hand_camera_speed_p90_m_s": 0.8557899467218826, "hand_world_valid_fraction": 0.4375, "hand_camera_valid_fraction": 0.4375, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.8557899467218826, "rham_p90_deg_s": null, "wms_p90": 0.998881262098843, "gt_coverage": 0.8732394366197181, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "f3e07ede5f2aff65e7145706c89cb7fce95316a463c057b71233dc57bc3b6533", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/d7c705b7c5512fb693d4709d/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f2d3095700aafb5c561e48993dd87879e71ece04de944d336e284df3e6782026", "mano_sha256": "125745c49e7944ff458f0b2ba83f4f2cfabd47aa3f2723714c3fc35c385bdec6", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "d81c8a918386842e1d24bf13", "sample_id": "d81c8a918386842e1d24bf13", "file": "short_originals_under_20s/DexYCB/d81c8a918386842e1d24bf13_MANO_GT_2.47s.mp4", "recording_id": "20200709-subject-01/20200709_142926", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 7/74 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 9 frames", "id": "d81c8a918386842e1d24bf13", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/d81c8a918386842e1d24bf13_MANO_GT_2.47s.mp4", "bytes": 851838, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.258205582190289, "p90": 1.0699111934464693, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "rhcm_continuous_estimate_m_s": { "median": 0.258205582190289, "p90": 1.0699111934464693, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "rhcm_discrete_m_s": { "median": 0.2582055821902895, "p90": 1.0699111934464698, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "wms": { "median": 0.9961420516499835, "p90": 0.9990660967183778, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 1.0699111934464693, "hand_camera_speed_p90_m_s": 1.0699111934464698, "hand_world_valid_fraction": 0.4527027027027027, "hand_camera_valid_fraction": 0.4527027027027027, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0699111934464698, "rham_p90_deg_s": null, "wms_p90": 0.9990660967183778, "gt_coverage": 0.9041095890410958, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "178cb5d68686247c7475725b962a227bd30180c2fc25885edaca7f1275abfca7", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/d81c8a918386842e1d24bf13/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f23f2c983b55f7079663af7136d979afac6d6cdbaf7d3ee45ac64f97c34b9172", "mano_sha256": "45fd71f35ca417fbaf49969a7db77cf254bf5e6b21bce14b9a3ba45f49e914de", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "dd7151b1b0cfa944a560e116", "sample_id": "dd7151b1b0cfa944a560e116", "file": "short_originals_under_20s/DexYCB/dd7151b1b0cfa944a560e116_MANO_GT_2.47s.mp4", "recording_id": "20200820-subject-03/20200820_142017", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 7/74 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 37 frames", "id": "dd7151b1b0cfa944a560e116", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/dd7151b1b0cfa944a560e116_MANO_GT_2.47s.mp4", "bytes": 841580, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.04303767777714593, "p90": 0.8770291186264043, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "rhcm_continuous_estimate_m_s": { "median": 0.04303767777714593, "p90": 0.8770291186264043, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "rhcm_discrete_m_s": { "median": 0.04303767777714593, "p90": 0.8770291186264043, "valid_pairs": 66, "valid_duration_s": 2.1999999999999997, "coverage": 0.9041095890410958 }, "wms": { "median": 0.9964743496029287, "p90": 0.9988772517573673, "valid_pairs": 47, "valid_duration_s": 1.566666666666667, "coverage": 0.6438356164383564 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.8770291186264043, "hand_camera_speed_p90_m_s": 0.8770291186264043, "hand_world_valid_fraction": 0.4527027027027027, "hand_camera_valid_fraction": 0.4527027027027027, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.8770291186264043, "rham_p90_deg_s": null, "wms_p90": 0.9988772517573673, "gt_coverage": 0.9041095890410958, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "99edcc96e193c6780ce990b355548f2de3d5dadb83646a71d729b2fbf40e68e2", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/dd7151b1b0cfa944a560e116/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "7b777721b9cfdaeee04b8a2f6100ab4bc05c254cfcbc9243bce5c4b0d3850ec8", "mano_sha256": "9d808f248aef9251a9920650c39749c2944c05abaef9bdac7bdd9c44220f4734", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "e4f406bdeba374f66b2140ef", "sample_id": "e4f406bdeba374f66b2140ef", "file": "short_originals_under_20s/DexYCB/e4f406bdeba374f66b2140ef_MANO_GT_2.40s.mp4", "recording_id": "20200918-subject-06/20200918_120812", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 3/72 output frames; may include genuine out-of-view frames; valid mesh not rasterized in 2 frames", "id": "e4f406bdeba374f66b2140ef", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/e4f406bdeba374f66b2140ef_MANO_GT_2.40s.mp4", "bytes": 916900, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.24196521842296725, "p90": 0.8857334960261011, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "rhcm_continuous_estimate_m_s": { "median": 0.24196521842296725, "p90": 0.8857334960261011, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "rhcm_discrete_m_s": { "median": 0.24196521842296673, "p90": 0.8857334960261019, "valid_pairs": 68, "valid_duration_s": 2.2666666666666666, "coverage": 0.9577464788732394 }, "wms": { "median": 0.9962735638609956, "p90": 0.9988816374572806, "valid_pairs": 61, "valid_duration_s": 2.0333333333333337, "coverage": 0.8591549295774649 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.8857334960261011, "hand_camera_speed_p90_m_s": 0.8857334960261019, "hand_world_valid_fraction": 0.4791666666666667, "hand_camera_valid_fraction": 0.4791666666666667, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.8857334960261019, "rham_p90_deg_s": null, "wms_p90": 0.9988816374572806, "gt_coverage": 0.9577464788732394, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "8dc086956ef6fe0d2e1eb5db9ff293fe5ab5523b9758dcc96e67958f0f3905d8", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/e4f406bdeba374f66b2140ef/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "8e8e5ded53025d87be31fb804a14e6bd8477f1e9f2bdec9eb2d7d1c596086901", "mano_sha256": "b2a3844562b70855dd6c52736a506eb34afd1a6cb84dab8db5f378d62025ee89", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "efd53853627fef08fe3dd6a1", "sample_id": "efd53853627fef08fe3dd6a1", "file": "short_originals_under_20s/DexYCB/efd53853627fef08fe3dd6a1_MANO_GT_2.47s.mp4", "recording_id": "20201022-subject-10/20201022_112939", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 19/74 output frames; may include genuine out-of-view frames", "id": "efd53853627fef08fe3dd6a1", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/efd53853627fef08fe3dd6a1_MANO_GT_2.47s.mp4", "bytes": 914448, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.29875032246201705, "p90": 1.0357993826749545, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "rhcm_continuous_estimate_m_s": { "median": 0.29875032246201705, "p90": 1.0357993826749545, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "rhcm_discrete_m_s": { "median": 0.29875032246201794, "p90": 1.0357993826749539, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "wms": { "median": 0.9967787152162859, "p90": 0.9990421776207459, "valid_pairs": 53, "valid_duration_s": 1.7666666666666666, "coverage": 0.726027397260274 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 1.0357993826749545, "hand_camera_speed_p90_m_s": 1.0357993826749539, "hand_world_valid_fraction": 0.3716216216216216, "hand_camera_valid_fraction": 0.3716216216216216, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 1.0357993826749539, "rham_p90_deg_s": null, "wms_p90": 0.9990421776207459, "gt_coverage": 0.7397260273972602, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "db5812f0ff7c6da5338ac231ca82c260ddac83ab6acc518296a249deb59a9b48", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/efd53853627fef08fe3dd6a1/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "f0eac08dc33d90b0f07460768b55d2483c93637585e87f3d37c60c871fd4c78a", "mano_sha256": "7a0670606f2d82d74b96c0cae50628059de215b822bcb152de439f887d9d2956", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "f4213e71dc0886ae498cfe88", "sample_id": "f4213e71dc0886ae498cfe88", "file": "short_originals_under_20s/DexYCB/f4213e71dc0886ae498cfe88_MANO_GT_2.47s.mp4", "recording_id": "20200918-subject-06/20200918_113645", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 5/74 output frames; may include genuine out-of-view frames", "id": "f4213e71dc0886ae498cfe88", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/f4213e71dc0886ae498cfe88_MANO_GT_2.47s.mp4", "bytes": 845679, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": 0.3239060040576245, "p90": 0.7490126178817655, "valid_pairs": 68, "valid_duration_s": 2.266666666666666, "coverage": 0.9315068493150683 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 68, "valid_duration_s": 2.266666666666666, "coverage": 0.9315068493150683 }, "rhcm_continuous_estimate_m_s": { "median": 0.3239060040576245, "p90": 0.7490126178817655, "valid_pairs": 68, "valid_duration_s": 2.266666666666666, "coverage": 0.9315068493150683 }, "rhcm_discrete_m_s": { "median": 0.3239060040576248, "p90": 0.749012617881765, "valid_pairs": 68, "valid_duration_s": 2.266666666666666, "coverage": 0.9315068493150683 }, "wms": { "median": 0.9973509735031669, "p90": 0.9986940993220469, "valid_pairs": 64, "valid_duration_s": 2.1333333333333337, "coverage": 0.8767123287671236 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "high", "relative_class": "high", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "high", "relative_class": "high", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.7490126178817655, "hand_camera_speed_p90_m_s": 0.749012617881765, "hand_world_valid_fraction": 0.46621621621621623, "hand_camera_valid_fraction": 0.46621621621621623, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.749012617881765, "rham_p90_deg_s": null, "wms_p90": 0.9986940993220469, "gt_coverage": 0.9315068493150683, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "fcca744b491652ee873293e92052c6f9f19c2f3559547f53abe31211703b38e6", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/f4213e71dc0886ae498cfe88/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "cd70ebf04fa5077000c638ac57e68dd74bca81d1be3b16935cc9d34d4f071f8e", "mano_sha256": "a9f56e769157e15be33bfe7e0a70caf856fe40d37c64e16636175424147ac279", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "f56c0ec751008f1cd25f46e5", "sample_id": "f56c0ec751008f1cd25f46e5", "file": "short_originals_under_20s/DexYCB/f56c0ec751008f1cd25f46e5_MANO_GT_2.40s.mp4", "recording_id": "20200908-subject-05/20200908_145353", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 9/72 output frames; may include genuine out-of-view frames", "id": "f56c0ec751008f1cd25f46e5", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/f56c0ec751008f1cd25f46e5_MANO_GT_2.40s.mp4", "bytes": 755827, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.2562451075289738, "p90": 0.7218320241278916, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "rhcm_continuous_estimate_m_s": { "median": 0.2562451075289738, "p90": 0.7218320241278916, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "rhcm_discrete_m_s": { "median": 0.2562451075289739, "p90": 0.7218320241278892, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "wms": { "median": 0.996111777080186, "p90": 0.9986165086737488, "valid_pairs": 62, "valid_duration_s": 2.0666666666666664, "coverage": 0.8732394366197181 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.7218320241278916, "hand_camera_speed_p90_m_s": 0.7218320241278892, "hand_world_valid_fraction": 0.4375, "hand_camera_valid_fraction": 0.4375, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.7218320241278892, "rham_p90_deg_s": null, "wms_p90": 0.9986165086737488, "gt_coverage": 0.8732394366197181, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "384582f1c68703947130d870deb4580a04b1c2591392a696d6963e5cc60b4347", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/f56c0ec751008f1cd25f46e5/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "3a8b58fc092f06b80d3e397f5249a0fe44ec576d67b44e2075fd8ac51e971e71", "mano_sha256": "1bed168bca66bfe6b6d1c668448b6059ee174c5f936071fff3236b571b5893f6", "duration_s": 2.4, "frames": 72, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "f822aac7c3262dbb86bee02a", "sample_id": "f822aac7c3262dbb86bee02a", "file": "short_originals_under_20s/DexYCB/f822aac7c3262dbb86bee02a_MANO_GT_2.40s.mp4", "recording_id": "20200928-subject-07/20200928_145010", "source_first_frame": "0", "source_last_frame": "71", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 17/72 output frames; may include genuine out-of-view frames", "id": "f822aac7c3262dbb86bee02a", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/f822aac7c3262dbb86bee02a_MANO_GT_2.40s.mp4", "bytes": 804274, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "camera_rotation": { "median": 3.9843289069162666e-13, "p90": 3.98432890691628e-13, "valid_pairs": 71, "valid_duration_s": 2.3666666666666667, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.13230574816017404, "p90": 0.8320346125404235, "valid_pairs": 54, "valid_duration_s": 1.8, "coverage": 0.7605633802816901 }, "cmh_m_s": { "median": 6.54644781534431e-15, "p90": 6.762232210575604e-15, "valid_pairs": 54, "valid_duration_s": 1.8, "coverage": 0.7605633802816901 }, "rhcm_continuous_estimate_m_s": { "median": 0.13230574816017188, "p90": 0.8320346125404254, "valid_pairs": 54, "valid_duration_s": 1.8, "coverage": 0.7605633802816901 }, "rhcm_discrete_m_s": { "median": 0.13230574690809038, "p90": 0.8320345563624015, "valid_pairs": 54, "valid_duration_s": 1.8, "coverage": 0.7605633802816901 }, "wms": { "median": 0.9924877517845727, "p90": 0.998782746764633, "valid_pairs": 54, "valid_duration_s": 1.8, "coverage": 0.7605633802816901 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 3.98432890691628e-13, "camera_valid_fraction": 1.0, "camera_pairs": 71, "hand_world_speed_p90_m_s": 0.8320346125404235, "hand_camera_speed_p90_m_s": 0.8320345563624015, "hand_world_valid_fraction": 0.3819444444444444, "hand_camera_valid_fraction": 0.3819444444444444, "cmh_p90_m_s": 6.762232210575604e-15, "rhcm_p90_m_s": 0.8320345563624015, "rham_p90_deg_s": null, "wms_p90": 0.998782746764633, "gt_coverage": 0.7605633802816901, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "77c35731df83c179d10b89dc3d714f095817fd1373dc927bb3fd9fea156823ae", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/f822aac7c3262dbb86bee02a/reference.npz", "window_s": [ 0.0, 2.3666666666666667 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 72, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } }, { "dataset": "DexYCB", "source_tag": "MANO_GT", "variant": "reference", "video_sha256": "dbb08789384dae98da79de654a0eb6522f541a0aadef006f4822233d50643624", "mano_sha256": "6797a385d9ad2d9e328c81aa19b4274e4523012902d5583f6d42479ef26fbb21", "duration_s": 2.467, "frames": 74, "fps": 30.0, "mano_source": "native DexYCB pose_m + subject mano calibration; official fitted annotation", "scene_depth_source": "native aligned sensor depths from eight synchronized calibrated DexYCB RGB-D views, millimetres to metres", "scene_camera_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "intrinsics_source": "source calibration", "scene_scale_anchor": "native depth / source coordinates", "manual_note": "", "previous_visual_review_hash_matched": "True", "technical_pass": "True", "training_label_decision": "pending_human_review", "usage": "candidate_review", "context_id": "fc6ac5e387cf2a25c529a314", "sample_id": "fc6ac5e387cf2a25c529a314", "file": "short_originals_under_20s/DexYCB/fc6ac5e387cf2a25c529a314_MANO_GT_2.47s.mp4", "recording_id": "20200928-subject-07/20200928_155118", "source_first_frame": "0", "source_last_frame": "73", "source_split": "official_s0_train", "training_label_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "quality_warnings": "no valid MANO_GT mesh in 19/74 output frames; may include genuine out-of-view frames", "id": "fc6ac5e387cf2a25c529a314", "aliases": [], "url": "https://huggingface.co/datasets/yangzijing/nine-source-hand-data-review-videos-20s/resolve/8db627faff1bd5bed598abd27f0bd83311193e50/short_originals_under_20s/DexYCB/fc6ac5e387cf2a25c529a314_MANO_GT_2.47s.mp4", "bytes": 885811, "width": 1282, "height": 508, "codec": "h264", "video_verified": true, "motion": { "metric_version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "parameter_sha256": "133630dda880c1a7efb87e7748ee723f75e9a83d8a65dd3ab30c8ccae18211d6", "training_allowed": false, "camera_translation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "camera_rotation": { "median": 0.0, "p90": 0.0, "valid_pairs": 73, "valid_duration_s": 2.433333333333333, "coverage": 1.0 }, "sides": { "left": { "whm_m_s": { "median": 0.13253596224301784, "p90": 0.7019785302555847, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "cmh_m_s": { "median": 0.0, "p90": 0.0, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "rhcm_continuous_estimate_m_s": { "median": 0.13253596224301784, "p90": 0.7019785302555847, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "rhcm_discrete_m_s": { "median": 0.1325359622430173, "p90": 0.7019785302555844, "valid_pairs": 54, "valid_duration_s": 1.7999999999999998, "coverage": 0.7397260273972602 }, "wms": { "median": 0.9935787497405286, "p90": 0.9987232808340738, "valid_pairs": 49, "valid_duration_s": 1.6333333333333337, "coverage": 0.671232876712329 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" }, "right": { "whm_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "cmh_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_continuous_estimate_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rhcm_discrete_m_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "wms": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "rham_deg_s": { "median": null, "p90": null, "valid_pairs": 0, "valid_duration_s": 0.0, "coverage": 0.0 }, "speed_spearman": null, "hand_class": "unknown", "relative_class": "unknown", "relative_pose_class": "unknown" } }, "hand_both_class": "unknown", "relative_both_class": "unknown", "relative_pose_class": "unknown", "camera_class": "stationary", "hand_class": "unknown", "relative_class": "unknown", "relative_definition": "wrist position relative to camera; orientation separately reported", "camera_translation_speed_p90_m_s": 0.0, "camera_angular_speed_p90_deg_s": 0.0, "camera_valid_fraction": 1.0, "camera_pairs": 73, "hand_world_speed_p90_m_s": 0.7019785302555847, "hand_camera_speed_p90_m_s": 0.7019785302555844, "hand_world_valid_fraction": 0.3716216216216216, "hand_camera_valid_fraction": 0.3716216216216216, "cmh_p90_m_s": 0.0, "rhcm_p90_m_s": 0.7019785302555844, "rham_p90_deg_s": null, "wms_p90": 0.9987232808340738, "gt_coverage": 0.7397260273972602, "hand_motion_definition": "wrist translation, not finger articulation", "status": "reference", "reason": "复用匹配源帧的参考轨迹；视频场景预测不参与计算。", "source_kind": "official_annotation_and_calibration", "camera_source": "official calibrated stationary external camera; inverse of world-to-camera extrinsics", "hand_source": "official per-camera fitted MANO joint_3d; not direct joint measurements", "reference_sha256": "0e5d03eb02efaf4432925892bb27ece6908500ab471cd946984adf2c9a1e9d6b", "metadata_sha256": "1915f03019042785dbd8951660e5c8d1e3f3a03259308df70eca1fe1bcc9cc82", "source_file": "long20_20260918/local/extra/previews/fc6ac5e387cf2a25c529a314/reference.npz", "window_s": [ 0.0, 2.433333333333333 ], "timestamps_source": "source frame IDs / nominal capture rate; not independently verified sensor PTS", "frame_count": 74, "common_world_residual_max_m": 0.0 }, "camera_display_source": "native stationary camera-to-world extrinsics, direction independently cross-checked across eight labeled views", "mano_display_kind": "官方 MANO GT（拟合标注）", "sequence_mapping": "source video_index.csv recording_id; cross-checked local case", "provenance": { "mano": { "label": "官方 GT（拟合）", "kind": "official", "detail": "数据集官方拟合 MANO 标注，仍可能存在拟合误差。" }, "pointcloud": { "label": "官方深度融合", "kind": "official", "detail": "由官方传感器实测深度经我们的过滤和融合生成；最终点云不是直接发布的官方点云 GT。" }, "camera": { "label": "官方固定标定", "kind": "official", "detail": "使用官方固定相机外参，并经跨视角一致性检查。" } } } ], "motion": { "version": "world_motion_v2.0", "parameters": { "epsilon_m_s": 0.001, "low_speed_gate_m_s": 0.02, "max_gap_factor": 1.5, "min_duration_s": 2.0, "min_coverage": 0.9, "stationary_translation_m_s": 0.02, "stationary_rotation_deg_s": 3.0, "hand_high_m_s": 0.3, "camera_high_m_s": 0.15, "camera_high_deg_s": 25.0, "relative_high_m_s": 0.3, "threshold_status": "provisional screening; sensitivity required, not calibrated sensor accuracy", "smoothing": "none", "aggregation": "any = maximum side P90; both requires both eligible sides" }, "code_sha256": "03ed0d9b94f0e661bae691eda690ef790908c20be067e25814077050f5618104", "scope": "reference trajectories only; no reconstructed scene inputs" }, "compatible_previous_revisions": [ "bf44a0f735c5c801b275e71f45cfe45742743a4f", "dee8bd5d02d7b7bebbace1e6b5dea6e24278714e" ], "pipeline_revision": "surface_fit_restore_20260921" }