Keypoint detection refers to a category of computer vision models that predict specific points (unlike the bounding boxes used in object detection). This technique is frequently employed in human pose estimation or finger tracking tasks, where the focus lies solely on an object's position rather than its size or shape.
