End-to-end vision-based fall detection benchmark and system using RTMPose, YOLO-Pose, ST-GCN++, CTR-GCN and ByteTrack with LOSO evaluation and sliding-window video inference.
computer-vision deep-learning pytorch human-pose-estimation action-recognition video-understanding fall-detection skeleton-based-action-recognition st-gcn healthcare-ai mmpose bytetrack yolo-pose rtmpose ctr-gcn
-
Updated
Sep 14, 2026 - Python