In this research, we are developing a system for estimating the 3D shape of an object from measurement information of only one viewpoint. While many methods for 3D reconstruction using deep learning have been proposed, this method attempts to do so using an RGB-D camera and a statistical shape model (SSM).