一种基于混合优化算法的医学图像配准方法

为了实现脑部多模医学图像配准，提出了一种基于混合优化算法的配准方法。该算法采用遗传算法中的杂交思想改进了混沌粒子群算法，并用最大互信息测度对脑部MRI及CT图像进行配准。该改进算法可有效地避免优化算子陷入局部极值，而且算法收敛快。实验结果证明了提出的基于遗传思想的改进混沌粒子群优化算法对多模医学图像配准具有有效性。

Medical image registration based on hybrid optimization algorithm

This paper proposed an approach of multimodal medical images registration based on CPSO (chaos particle swarm optimization) algorithm improved by GA (genetic algorithm). And applied MMI to the registration. The method had many advantages such as avoiding optimizer trap into local maxima, quick convergence. The matching results of MRI /CT demonstrate that registration dependent on CPSO algorithm improved by GA is a robust, fast registration method.