nodes/ahkimkoo-ComfyUI-VideoActorExtract
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Commit Graph

  • 4912be31f3 feat: bump version to 1.2.0 main fix/identity-temporal-gap-penalty Cherokee 2026-05-06 14:30:34 +08:00
  • 40bb1fa1ee feat: bump version to 1.1.0, add GitHub Actions workflow for Comfy Registry publishing Cherokee 2026-05-06 13:33:04 +08:00
  • a883264d0d chore: update .gitignore and workflow JSON format Cherokee 2026-05-06 13:27:41 +08:00
  • 99ea23823b fix: UnboundLocalError in RGBA encoding loop — composited must be assigned in all branches Cherokee 2026-05-06 13:07:39 +08:00
  • 4796bf9026 feat: add back transparent option — PNG previews + WebM video Cherokee 2026-04-30 18:13:19 +08:00
  • 28b3cda552 improve: enable CUDA GPU for InsightFace on Linux + RTX 4090 Cherokee 2026-04-30 18:08:10 +08:00
  • 91c4f73130 improve: reduce InsightFace sampling from 30 to 10 frames, single crop attempt Cherokee 2026-04-30 17:39:08 +08:00
  • 42ba415e2c fix: interpolated frames reuse previous composited frame, no redundant copy Cherokee 2026-04-30 17:12:31 +08:00
  • 1ca35003ca docs: update README — remove transparent, keep JPG/MP4 only Cherokee 2026-04-30 17:03:09 +08:00
  • 207fdae844 revert: remove transparent option, keep JPG/MP4 for performance Cherokee 2026-04-30 17:02:19 +08:00
  • ba5aed299e fix: previews always output as PNG with alpha channel Cherokee 2026-04-30 16:37:27 +08:00
  • 85e2dc2fd2 feat: add bg_color option for cutout background (transparent/green/blue/black/white) Cherokee 2026-04-30 16:35:17 +08:00
  • 2f58b7e589 fix: exclude frames without face detection from preview candidates Cherokee 2026-04-30 15:54:42 +08:00
  • 3f91b426de fix: sort preview images by face area instead of mask bbox area Cherokee 2026-04-30 15:22:23 +08:00
  • 322f8d9b71 improve: performance optimization — downscale inference, defer green screen, add skip_every_n Cherokee 2026-04-30 14:04:35 +08:00
  • 7f9d51c55d chore: prepare pyproject.toml for ComfyUI registry publishing Cherokee 2026-04-30 11:44:28 +08:00
  • 0abc9b54c3 docs: add plain-language explanation for face_threshold and min_track_length Cherokee 2026-04-30 11:32:35 +08:00
  • 2092da6623 feat: auto-check and download missing models at pipeline start Cherokee 2026-04-30 11:06:14 +08:00
  • 7b19351c36 feat: preview sorting by bbox area, output frame indexes as INT list, update README Cherokee 2026-04-30 10:59:17 +08:00
  • 59e768891c improve: SelectActorPreview returns full IMAGE batch instead of single photo Cherokee 2026-04-29 18:19:08 +08:00
  • d6d208081b feat: add SelectActorPreview node, identity temporal constraints, and preview images to disk Cherokee 2026-04-29 09:34:33 +08:00
  • f1cd44a280 docs: rewrite installation section for ComfyUI Manager distribution Cherokee 2026-04-28 17:45:28 +08:00
  • 75b7a432f5 feat(preview): reshape actor_preview_images to [num_actors, 5, H, W, 3] Cherokee 2026-04-28 16:16:09 +08:00
  • 1701944636 fix: identity clustering with track splitting, spatial tiebreaker, and temporal constraints Cherokee 2026-04-28 15:39:18 +08:00
  • 63e4f0e525 docs: comprehensive README with architecture, usage, and testing guide Cherokee 2026-04-24 16:08:32 +08:00
  • 764c3ed9b3 improve: identity clustering with connected components, better face detection, and debug logging Cherokee 2026-04-24 15:48:33 +08:00
  • 3bc44d42dd fix: correct identity renumbering to preserve integer track ID keys Cherokee 2026-04-24 10:19:04 +08:00
  • aa2e07167d fix: per-frame processing to avoid 14GB memory spike Cherokee 2026-04-23 18:01:17 +08:00
  • 63fc34bd5a fix: improve identity clustering + model paths to ComfyUI/models/video-actor-extract Cherokee 2026-04-23 15:38:30 +08:00
  • 967fe6da2b fix: mask tracker for complete person extraction + continuous segments Cherokee 2026-04-23 10:52:20 +08:00
  • df3feddbb1 fix: use unique output directory per run (uuid4 short) Cherokee 2026-04-22 18:32:24 +08:00
  • eef072e85e feat: integrate PersonSegmenter for mask-based green screen extraction Cherokee 2026-04-22 17:08:54 +08:00
  • 713596c1d3 fix: remove incorrect squeeze(0) on 4D IMAGE tensor Cherokee 2026-04-21 19:16:06 +08:00
  • 4a819d659e feat: update node to accept IMAGE input from VHS LoadVideo Cherokee 2026-04-21 19:09:32 +08:00
  • 41510eecbe feat: add ComfyUI workflow file and usage examples Cherokee 2026-04-21 18:54:20 +08:00
  • ebad5ec8a2 fix: resolve namespace conflict with ComfyUI nodes module Cherokee 2026-04-21 17:34:14 +08:00
  • 4c04bcfc26 test: full pipeline test passed Cherokee 2026-04-21 17:21:03 +08:00
  • 67fdc05c9c fix: import DEFAULT_YOLO_MODEL in actor_extractor node Cherokee 2026-04-20 15:01:07 +08:00
  • 53b2246c81 chore: add .gitignore, remove pycache files Cherokee 2026-04-20 15:00:04 +08:00
  • 6e6564ff6a feat: add model_path parameter and README Cherokee 2026-04-20 14:59:51 +08:00
  • 968b9e6b19 feat: initial implementation of VideoActorExtract plugin Cherokee 2026-04-20 14:36:09 +08:00
  • e88b73dafd Initial commit: requirements and architecture docs Cherokee 2026-04-20 14:09:02 +08:00