ComfyUI Node
Animal Pose and Detection
Detects animal poses from images using ViTPose (AP10k/APT36k) and YOLO. Optionally retargets poses based on a reference image.
Animal Pose and Detection
- model
- images
- retarget_image
- pose_data
- key_frame_body_points
- bboxes
◄width832►
◄height480►
CategoryWanAnimalPreprocess
Inputs (5)
| Name | Type | Default | Description |
|---|---|---|---|
| model | POSEMODEL | From ONNX Animal Detection Model Loader. | |
| images | IMAGE | Input video frames as IMAGE batch. | |
| width | INT | 83264–2048 | Width of the generation |
| height | INT | 48064–2048 | Height of the generation |
| retarget_imageopt | IMAGE | Optional reference image for pose retargeting |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| pose_data | POSEDATA | Animal pose dict bundle. Feed into Draw Animal ViT Pose. |
| key_frame_body_points | STRING | Key-frame body keypoints as JSON string. |
| bboxes | BBOX | Per-frame body BBOX list. |