Finding Deformable Shapes by Point Set Matching Through Nonparametric Belief Propagation

摘要

This paper addresses the problem of finding a deformable shape by matching a point distribution model to the observation. A probabilistic graphical model is built for the point distribution model. The point correspondence and optimal model parameters are found by carrying out nonparametric belief propagation on the graphical model. Experiments on a point distribution model of the proximal model verified the idea.