Medical Image Registration by Maximization of Combined Mutual Information and Edge Correlative Deviation

A new approach of Medical Image Registration based on theories of mutual information and edge correlative deviation has been given in this paper. It takes advantage of the information contributed by the overall intensity values of original images as well as the correlation derived from the voxels' positions in edge images. It has been proven by tests that this new method inherits most merits of the former approaches and additionally bears several improved attributes: 1) The climaxes of parameters' curves are more obvious; 2) The errors are more diminished in extreme conditions where images lack intensity values; 3) It is more robust in resistance to aggravated white noise.