基于仿射变换与B样条自由形变的医学图像配准

摘要

In medical image registration process,human tissues and organs are complex,the traditional bspline registration is difficult to describe the deformation of the tissues and organs,and the results of registration have the error.In this paper,the mutual information is used as the measurement of similarity,the affine transformation is used as global rough registration,as well as b-spline free-form deformation as local registration and the multi-resolution strategy is introduced to improve the accuracy of registration,reduce the computational complexity and complete the registration of brain CT and MR1 medical image.The experiment results show that the method of registration can complete the registration of multi-scale deformation with prospective effect.%针对医学图像配准的人体组织器官的几何形体复杂多样,传统B样条配准方法较难很好地描述组织器官形变,配准结果有较大的误差等问题,使用互信息作为相似性测度,采用仿射变换作为全局配准的粗配准以及B样条自由形变作为局部配准的精确配准,为了提高配准精确度以及降低计算复杂度,引入多分辨率策略,对脑部CT和MRI图像进行配准.实验结果表明,该方法对多尺度的形变图像都有很好的配准效果.