Dense Matching Based on Subspace Learning for Non-Rigid Object
Qian Zhang · 2016
A dense matching method for non-rigid object is proposed in the paper. Our method is inspired by the continuous convex surface of non-rigid object. The similarity of each patch in a surface structure optimizes an energy minimization through Graph cut model. In order to improve the efficiency of method, a subspace learning algorithm is proposed with robust discrete matched points. The Evaluation of our method is given to guarantee the effectiveness of proposed research.