Common CV use cases such as human pose estimation require keypoints in a “skeleton”/“instance” format, where a set number of keypoints are grouped together as a single instance. For example, consider the COCO people dataset. This dataset contains 17 distinct keypoints which each instance of a person for things like right eye, left eye, right elbow, left elbow, etc.
The current KeyPointLabels, as far as I know, are just meant to represent a single keypoint. There’s no representation for a skeleton or grouping of keypoints, often in a graph representation for visualization purposes. Other annotation platforms like CVAT clearly support this. Am I missing how LabelStudio supports this? Or if they do not, what is the recommended workaround for annotating this type of data?