A new image registration algorithm based maximum mutual information

A new image registration algorithm which based on immune optimal algorithm had been proposed in this paper. The mutual information was chosen to direct the registration process. Because that the mutual information was maximum when the original images were strictly alignment, so the registration process could be turned to the problem how to find the optimal geometric transformation parameters of yaw angle and coordinate offset between the original images. The immune optimal algorithm was adopted to calculate the optimal yaw angle and coordinate offset. It took the mutual information as the affinity function of the immune optimal algorithm, the yaw angle and coordinate offset were taken as antibodies of the algorithm. The mutual information reached to maximum when it obtained the optimal geometric transformation parameters. The image registration results had shown that the algorithm could easily to stable and convergence. It could solve the image registration problem perfectly.