Viser: drop hand skeleton, route wrist trail to MANO mesh vertex (105 clips)

Upload 105 clip(s) (4.85 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-22T08:46:54+00:00] (13/105)

Upload 105 clip(s) (4.85 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-22T07:52:33+00:00] (13/105)

P0: remove 11 TRUE_DUP phantom oids (9 clips) (2/9)

SAM3D canonical-init-frame fix: 12 clips refreshed (sanity 5 + batch 7) (3/12)

Upload 105 clip(s) (4.88 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-19T15:49:37+00:00] (13/105)

Upload 105 clip(s) (4.88 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-19T10:01:54+00:00] (13/105)

Upload 105 clip(s) (4.88 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-19T09:13:29+00:00] (13/105)

Upload 105 clip(s) (4.88 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-19T07:26:37+00:00] (13/105)

Upload 105 clip(s) (4.88 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-19T06:12:25+00:00] (13/105)

Upload 105 clip(s) (4.88 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-19T04:38:36+00:00] (13/105)

Upload 105 clip(s) (4.97 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-19T03:41:50+00:00] (13/105)

Upload 105 clip(s) (4.97 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-18T07:24:56+00:00] (13/105)

Upload 105 clip(s) (4.97 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-17T18:46:58+00:00] (13/105)

Upload 105 clip(s) (4.97 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-17T17:24:47+00:00] (13/105)

Upload 105 clip(s) (4.97 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-17T16:46:06+00:00] (13/105)

2026-05-17: drop oid 3 'glass' — duplicate of blender (mask containment 1.0)

2026-05-17: orphan sam3_obj_data cleanup — drop per-frame mask/OBB/pose_t for removed oids (fixes ghost outlines) (4/26)

2026-05-17: spurious_filter medium mode — drop 26 oids that are far from hand zone AND static (4/17)

2026-05-15: state machine v2 — STATIC / GRASPED_L/R/BOTH / MOVING + L/R-aware grasp (13/105)

Upload 105 clip(s) (5.04 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-13T18:29:43+00:00]

Upload 105 clip(s) (5.04 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-13T17:54:01+00:00]

Upload 105 clip(s) (5.04 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-13T11:00:21+00:00]

D-track refresh after A100 SAM3D recovery — -4eiU1SHpzY_111.9_123.8

Upload 105 clip(s) (4.75 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-13T06:42:45+00:00]

Upload 105 clip(s) (4.70 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+102 more) [2026-05-13T05:49:53+00:00]

Upload 112 clip(s) (4.96 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+109 more) [2026-05-13T02:21:54+00:00]

Upload 112 clip(s) (4.99 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+109 more) [2026-05-12T22:11:25+00:00] (13/112)

Upload 112 clip(s) (4.99 GB) — --oo8_XIuOM_799.5_809.8, --oo8_XIuOM_900.3_917.4, -0RheyDV3a0_474.8_487.3, ... (+109 more) [2026-05-12T20:49:27+00:00] (13/112)

v1.11: pure canonical_scale (hf_reference simplified). No ICP / Umeyama / sanity refinement. Mesh scale = SAM3D canonical_scale directly; R = canonical_rotation_quat; per-frame translation = obs centroid + Gaussian smooth.

v1.10: restore original 5-stage tracking (estimate_anchor_pose FGR+ICP+Umeyama + track_6dof flow propagation). 40/225 oids get full 6DoF tracking; rest fall back to centroid translation + identity rotation. Scale closer to A100 reference (cereal box 0.219 vs reference 0.263).

v1.9: scale_correction is now ABSOLUTE (PLY units→world m, matches A100 reference convention). Robust min-enclose algorithm: per-axis [2,98]% pre-filter + median over axes of per-axis 90% enclose (depth-fringe-resistant)

v1.8: tracking uses SAM3D canonical_rotation_quat (per-clip initial orientation) + min-scale-95% enclose algorithm; object_pose.bin now reads pose_track_info.T_seq (full 6DoF tracked pose, not stale PCA OBB)

v1.7: SAM3D refill round 2 (226 PLYs, +25 from v1.6) — decoder cleanup frees 1.7GB GPU; native-res bbox crop (SAM3D internally resizes to 518)

v1.5: pose tracking (position-first) — SAM3D mesh follows SAM2 mask centroid frame-by-frame; rotation fixed at anchor; 164/164 oid tracked

v1.4: OBB Level 1 fix — mask erosion (2px) + MAD outlier filter, eliminates depth-fringe contamination causing diagonal OBBs

v1.3: SAM3D meshes (164 PLYs, ~383 MB) + viser scenes with mesh layer enabled (Plan B mask scoring: area * (1-hand_iou) * static_ratio)

viser 1.0.27 — fix empty 3D viewer (client/server format mismatch)

v1.2: exact WiLoR joints_2d (re-extracted) + per-clip viser 3D recordings (object pointcloud+OBB+state-dot, no SAM3D mesh)

v1.1: 105 clips, SAM3-quality-sorted index (best→worst), WiLoR-native skeleton overlay

WiLoR-native joints_2d for skeleton overlay (88 clips re-exported, MoGe-2-independent 2D projection bypasses focal estimation errors)

Fix trim consistency: scene.json fps/duration use trim window (was 2x bug), video_source pinned to trim, frontend trim review pre-loads saved values

Step 1 full re-run @ 15fps with compound+fallback prompts and containment merge (55 clips)

Re-export 21 clips: scene.json objects/[] now populated from sam3_obj_data when no SAM3D mesh; silver-can clip uses 'silver can' instead of 'can of olives' (SAM3 hits 4/4 vs 3/4)

Replace samples/ with 21-clip Step 1 test batch (no SAM3D, no 6DoF)

recording.viser: include sam3d object meshes (fix stale ply_path); README: viewer:false + snapshot_download docs

Refresh: 48 clips with full descriptions + viser recordings

Refresh scene.json: surface Action100M action_detailed / summary / video_title; sync index.json

Refresh dataset: 48 trim-reviewed clips, manual SAM3 prompts, full descriptions in scene.json

Switch mask.mp4 to RGB cutout of SAM-tracked objects