出版社：Academy & Industry Research Collaboration Center (AIRCC)
摘要：This paper presents a relaxation labeling technique with newly defined compatibility measuresfor solving a general non-rigid point matching problem. In the literature, there exists a pointmatching method using relaxation labeling, however, the compatibility coefficients always takea binary value zero or one depending on whether a point and a neighboring point havecorresponding points. Our approach generalizes this relaxation labeling approach. Thecompatibility coefficients take n-discrete values which measures the correlation between edges.We use log-polar diagram to compute correlations. Through simulations, we show that thistopology preserving relaxation method improves the matching performance significantlycompared to other state-of-the-art algorithms such as shape context, thin plate spline-robustpoint matching, robust point matching by preserving local neighborhood structures andcoherent point drift.
