Human pose estimation models.
Represents a single keypoint in a pose.
Represents a detected human pose containing multiple keypoints.