Learning to Estimate 3D Human Pose from Point Cloud

3D pose estimation is a challenging problem in computer vision. Most of the existing neural-network-based approaches address color or depth images through convolution networks (CNNs). In this paper, we study task human from images. Different CNN-based method, propose deep network for by taking point cloud data as input to model surface complex structures. We first cast 2D clouds and directly predict joint position. Our experiments on two public datasets show that our approach achieves higher accuracy than previous state-of-art methods. The reported results both ITOP EVAL demonstrate effectiveness method targeted tasks.