Lei Wang, Linlin Ge, Shan Luo +3
Structure-from-Motion (SfM) aims to recover 3D scene structures and camera poses based on the correspondences between input images, and thus the ambiguity caused by duplicate struc…