A factorization-based projective reconstruction algorithm with circular motion constraint

In this paper, we propose a projective reconstruction algorithm for a circular motion image sequence. We first formulate the circular motion constraint in the Euclidean frame, and then deduce its expression in a projective frame. The circular motion constraint is gradually enforced during the iterations of a projective reconstruction. This approach can be used to deal with both constant and varying intrinsic parameters. Experimental results for synthetic and real data are presented to illustrate the performance and improvements of our approach over methods based on general motion.