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] (38/112)

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] (38/112)

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