nodes/ahkimkoo-ComfyUI-VideoActorExtract
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
14 Commits 2 Branches 0 Tags
63fc34bd5a6b4e8a534fa9528d3d9ecb9d2d12fd
Commit Graph
14 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Cherokee 63fc34bd5a fix: improve identity clustering + model paths to ComfyUI/models/video-actor-extract 2026-04-23 15:38:30 +08:00
Cherokee 967fe6da2b fix: mask tracker for complete person extraction + continuous segments 
- Replace ByteTrack mask assignment with centroid-distance MaskTracker
- Every detected person mask gets extracted, no more missed people
- Continuous segment output with interpolation for small gaps
- Accurate start_frame/end_frame/frame_count in JSON
 2026-04-23 10:52:20 +08:00
Cherokee df3feddbb1 fix: use unique output directory per run (uuid4 short) 2026-04-22 18:32:24 +08:00
Cherokee eef072e85e feat: integrate PersonSegmenter for mask-based green screen extraction 
- Replace bbox crop with YOLOv8-seg person segmentation
- Full-frame green screen: only person silhouette kept
- Top 5 preview images per actor by mask area (IMAGE output)
- Accurate time segmentation: only frames where person is visible
- No green gaps between segments
- Add seg_model_path parameter
 2026-04-22 17:08:54 +08:00
Cherokee 713596c1d3 fix: remove incorrect squeeze(0) on 4D IMAGE tensor 2026-04-21 19:16:06 +08:00
Cherokee 4a819d659e feat: update node to accept IMAGE input from VHS LoadVideo 
- Changed video_path (STRING) to images (IMAGE) as required input
- Removed fps_sample parameter (process all frames from VHS)
- Added optional video_path for metadata fallback
- Updated workflow file with VHS_LoadVideo node
- Saved workflow to ~/App/ComfyUI/user/default/workflows/
- Node now connects: VHS LoadVideo IMAGE -> VideoActorExtractor
 2026-04-21 19:09:32 +08:00
Cherokee 41510eecbe feat: add ComfyUI workflow file and usage examples 
- video-actor-extract.json: drag-and-drop workflow for ComfyUI
- Includes VideoActorExtractor + Save Text File nodes
- README with parameter tuning guide
- Saved to both examples/ and ComfyUI output/workflows/
 2026-04-21 18:54:20 +08:00
Cherokee ebad5ec8a2 fix: resolve namespace conflict with ComfyUI nodes module 
- Rewrote __init__.py to use importlib for loading modules
- Avoids conflict between plugin's nodes/ dir and ComfyUI's nodes.py
- Tested in ComfyUI with 1080x1920 18.7s video
- Successfully detected 10 actors with green screen output
- actor_1 correctly shows 2 segments (appeared mid-video, left, returned)
- All actor_info.json and per-actor MP4s generated correctly
 2026-04-21 17:34:14 +08:00
Cherokee 4c04bcfc26 test: full pipeline test passed 
- YOLOv8 model downloaded successfully (6.5MB)
- Tested with sample_demo_1.mp4 (640x360, 25fps, 9.7s)
- Detected 1 actor (actor_0)
- Output: actor_0.mp4 (497KB, green screen cropped)
- Output: actor_info.json (valid JSON with frame ranges)
- Node loads correctly in ComfyUI environment
- Symlink to custom_nodes verified
 2026-04-21 17:21:03 +08:00
Cherokee 67fdc05c9c fix: import DEFAULT_YOLO_MODEL in actor_extractor node 2026-04-20 15:01:07 +08:00
Cherokee 53b2246c81 chore: add .gitignore, remove pycache files 2026-04-20 15:00:04 +08:00
Cherokee 6e6564ff6a feat: add model_path parameter and README 
- Added model_path input to VideoActorExtractor node
- Added comprehensive README with setup instructions
- Cleaned up test artifacts
 2026-04-20 14:59:51 +08:00
Cherokee 968b9e6b19 feat: initial implementation of VideoActorExtract plugin 
- YOLOv8 person detection (ultralytics)
- ByteTrack multi-object tracking (with greedy fallback)
- InsightFace face recognition + cosine similarity clustering
- Green screen actor cropping
- Multi-segment video merging with ffmpeg
- ComfyUI node: VideoActorExtractor (category: video/actor)
- Output: per-actor MP4 + actor_info.json
 2026-04-20 14:36:09 +08:00
Cherokee e88b73dafd Initial commit: requirements and architecture docs 2026-04-20 14:09:02 +08:00