This workflow performs multi-person human pose detection on video using the SDPose-OOD whole-body model. It ingests a video with LoadVideo, extracts frame size and FPS via GetVideoComponents, and feeds each frame into an SDPose subgraph (node ID 01b6a731-fb78-4070-9a38-c87146da9604). Inside the subgraph, an object detector proposes person bounding boxes per frame and the SDPose whole-body checkpoint (sdpose_wholebody_fp16.safetensors) estimates keypoints for body, hands, and face. The PrimitiveInt node controls max_detections, letting you cap how many people are processed per frame in crowded scenes.

The workflow outputs two visualizations: bounding boxes via DrawBBoxes and a pose map overlay blended onto the original frames with ImageBlend. ResizeImageMaskNode ensures the pose map matches the source frame dimensions, and PreviewImage offers a quick sanity check before exporting. Finally, CreateVideo reassembles the annotated frames at the original FPS, and SaveVideo writes the result. The SDPose logic is packaged as a subgraph, so you can right-click to Unpack Subgraph and tweak internal thresholds or routing if you want finer control.