OpenPose

An open-source computer vision library developed by Carnegie Mellon University that detects human body, hand, facial, and foot keypoints in real-time from images or video. OpenPose extracts 25 body keypoints, 21 keypoints per hand, and 70 facial landmarks, providing a skeletal representation of human pose. In accessibility research, OpenPose is widely used for sign language recognition and translation because it captures the hand shapes, body positions, and facial expressions essential to sign language grammar. Its pose-based features are robust to variations in camera angle, signer position, and lighting conditions, making it valuable for real-world sign language processing systems.