Abstract: Point clouds captured from 3-D scanners are often sparse and incomplete due to occlusions, limited viewpoints, and sensor constraints. These limitations hinder applications in robotics, ...