Vision based robot manipulation uses cameras to capture one or more images of a scene containing the objects to be manipulated. Taking multiple images can help if any object is occluded from one viewpoint but more visible from another viewpoint. However, the camera has to be moved to a sequence of suitable positions for capturing multiple images, which requires time and may not always be possible, for example due to reachability constraints. Estimating complete geometries of objects by leveraging 3D shape completion can prove advantageous in computing grasps of an object from just one viewpoint. In this project, I want to further improve current state-of-the-art on shape completion for robot manipulation.