基于改进微粒群优化算法的互信息医学图像配准

医学图像配准是寻找使两幅图像对应点达到空间位置和解剖结构上一致的过程，对于医学临床科研、诊断等方面具有非常重要的意义。基于互信息的配准方法是目前医学图像配准中无创、自动化程度高且配准精度很高的一种方法，已被广泛应用。但是由于插值赝像导致在其目标函数中存在幅值振荡现象，使用局部最优化搜索有时会终止于局部极值，得到错误的配准参数。提出针对互信息配准方法的特点，使用改进微粒群全局搜索算法，调节该算法的参数以适应不同搜索阶段，保证了最优化搜索的准确性，提高了基于互信息方法配准的成功率。

Mutual Information-Based Medical Image Registration Method with Improved Particle Swarm Optimization Algorithm

Medical image registration is the process of finding a geometric transformation between the respective imagebased coordinate systems that maps a point in the first image to the point in the second one that has the same patient-based coordinates and the same anatomic location, so it has significant application in diagnosis and clinical scientific research. The image registration method based on mutual information has been accepted as the most accurate automated and efficient method. But there are many fluctuations in the registration function resulting from interpolation artifacts, which hinder the local optimization procedure and lead to convergence at local optimum even wrong registration. In this paper, a new optimization method using improved particle swarm global search algorithm is presented, which adjusts the parameters in different search phrases. This method is validated to ensure the accuracy of optimization search and improve the correctness of the registration process based on mutual information.