Find Point Correspondences on Contour between Template Image and its Target with SVD and Euclidean Distance
A new algorithm is presented in this paper to determinate the point correspondences on contour between template image and its target after affine transformation. In the algorithm, the singular value decomposition(SVD)is applied to the contour point sets of the template and target image respectively for eliminating the influences of the shear and scale in the affine transformation. The Euclidean distance between the contour point and the center of the shape are taken as the feature to form the reference sequence and comparative sequences, and then grey relational analysis (GRA) is used to find the best correlation sequence. After two contour sequences with the best correlation are found, the corresponding points between the two contours can be decided also. Finally the affine transformation parameter can be calculated and image matching can be realized by this way. Compared with the similar methods, experiments show that the proposed method has lower computational complexity and better accurate for image matching.