An adaptive simulated annealing scheme for multi-modality medical image registration by maximization of mutual information

Mutual information (Ml) is a widely used similarity metric for multimodality image registration. However, its robustness is affected by existence of local maxima which make it difficult for numerical optimizers to find the optimal transformation. In this paper, we proposed an adaptive simulated annealing scheme for optimizing the mutual information in a coarse-to-fine multi-resolution approach. The proposed search strategy is based on using a non-stochastic optimizer in primary stages and the simulated annealing in the final stage. The experimental results for volume images of CT, MR and PET are presented