Upload 82 clip(s) (3.71 GB) — -821U1ZYZOs_245.0_249.0, -821U1ZYZOs_321.4_325.4, -821U1ZYZOs_326.1_330.3, ... (+79 more) [2026-05-22T09:07:28+00:00] (24/82)

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] (47/105)

P4: remove 6 likely-DUP AMBIG phantom oids (5 clips, heuristic d3d<0.7m+size_ratio>0.5) (3/5)

P0: remove 11 TRUE_DUP phantom oids (9 clips) (5/9)

SAM3D hand_iou-gate rerun: 26 clips refreshed (D-track + filters + export) (13/26)

Upload 82 clip(s) (3.72 GB) — -7uLbtVsmQs_874.8_879.1, -7uLbtVsmQs_874.8_879.9, -821U1ZYZOs_245.0_249.0, ... (+79 more) [2026-05-19T10:23:31+00:00] (26/82)

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] (49/105)

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] (49/105)

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] (49/105)

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] (49/105)

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] (49/105)

Upload 64 clip(s) (3.00 GB) — -7uLbtVsmQs_874.8_879.9, -821U1ZYZOs_245.0_249.0, -821U1ZYZOs_321.4_325.4, ... (+61 more) [2026-05-17T18:02:06+00:00] (26/64)

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] (49/105)

2026-05-17: spurious filter v2 — same-prompt duplicate detection (Pour drink 3 background whiskey bottles + 2 others) (2/3)

2026-05-15: refresh 11 mask-drift clips with hand-proximity init_frame penalty (5/11)

2026-05-15: refresh 22 mask-drift clips (Q1×2 outlier penalty + new init_frames + D-track) (13/22)

D-track refresh after A100 SAM3D recovery — -EI-_9BR7j8_234.0_282.9

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] (49/112)

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)

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

Add batch: 88 clips total @15fps with curated objects