为了提高医学图像配准的效率,提出一种基于并行计算和多层次B样条函数的图像配准方法。该方法提取图像均匀内部点及轮廓点作为参考图像特征点,采用并行计算技术在浮动图像上选择特征点对;在形变过程采用B样条曲面函数插值来完成,通过多层次B样条函数逐步优化形变效果。实验结果表明,该方法适合于肺器官CT-PET图像快速配准。
In order to improve the efficiency of medical image registration, an image registration method based on hierarchical B-spline and parallel computing was proposed in this paper. First, landmarks were selected from the evenly distributed inner points and the outer contours of the images, and the correspondence points were automatically determined by using the parallel computing technology. Then the registration deformation was accomplished by hierarchical B-splines algorithm. Experimental results demonstrate that the method is efficient.