Skeleton Keypoints Extraction Method Combined with Object Detection

Big data is a comprehensive result of the development Internet Things and information systems. Computer vision requires lot as basis for research. Because skeleton can adapt well to dynamic environment complex background, it used in action recognition tasks. In recent years, skeleton-based has received more attention field computer vision. Therefore, keypoints human skeletons are essential describing pose estimation predicting human. This paper proposes point extraction method combined with object detection, which focus on keypoints. After large number experiments, our model be detection points extraction, efficiency improved.