Abstract: Medical image registration technique is floating medical image and the reality of physical space match point, which helps doctors to quickly find the lesion area. In the image registration optimization algorithm, if the optimal value is always the same, it will be obtained locally optimal solution. Therefore, we propose optimization algorithm based on mutual information and mixed medical image registration algorithm, and improves the ability of global optimization based on this technology. Experimental results show that it has better image registration accuracy, and has robustness on salt and pepper noise and Gaussian noise.

Key words:  medical image registration, mutual information, particle swarm optimization (PSO)