基于改进粒子群优化算法的互信息图像配准

摘要

In order to realize the fast precise image registration, a mutual information image registration method based on improved particle swarm optimization is proposed. Mutual information is used as the similar measure. The improved particle swarm optimization algorithm solves the space transformation parameters of registration. The organization concept is introduced in the methtod to divide the entire population into several sub-groups, meanwhile the mutation operation in genetic algorithm is introduced to reduce the local extreme. And improved PSO is used in the medical image registration. The results show that the method can achieve more satisfied results of the image registration.%为了实现快速精确的图像配准,提出了基于改进粒子群优化算法的互信息图像配准方法,以互信息作为图像配准的相似性测度,使用改进的PSO算法来求解配准所需的空间变换参数.改进的粒子群算法引入组织的概念把整个种群划分为多个子群体共同进化,并引入变异运算减少算法陷入局部最优.把改进的粒子群优化算法应用到医学图像配准领域上来,实验结果表明,算法能够得到比较满意的配准结果.