pose keypoint

A specific, labeled body-joint location (e.g. left elbow, right knee) detected in an image or frame.

A pose keypoint is a specific, labeled anatomical location -- a joint such as a wrist, elbow, or knee -- that a pose-estimation model detects in an image. A full set of detected keypoints, connected in the anatomically expected order, forms a skeleton; action recognition applications are typically built on a sequence of skeletons across consecutive frames.

Examples

A pose model reports a person's left wrist at pixel $(140,220)$ and their right knee at $(210,480)$ , connecting the full set into a skeleton.