The feature-based image registration method is common used technique in the image registration. It is with strong validity and reliability. Firstly, the method to extract the shape feature point is proposed in the paper, which unites GVF Snake and curve evolution. The method is with better robustness in anti-noise and cutting redundancy. Secondly, the content of shape matching in shape registration is expanded and the application of the shape similarity in shape matching is designed reasonable. Finally, a forecast is found in the research that the Particle Swarm Optimization (PSO) can be used in image registration. An improved PSO called Adaptive Fuzzy Discrete PSO is applied in field of image registration. It offers a new way to study the image registration for the future.As is issued above, a new image registration based on target shape feature is proposed in this paper. Favorable image registration effects are achieved by experiments. |