提出了一种基于轮廓的多设备医学图象的刚性配准算法。在半自动轮廓搜索的基础上，运用迭代最近邻点方法搜索轮廓间的对应点，最小化代价函数并获得最佳刚性变换，实现图象的精确配准。通过多次选择初始变换的搜索策略，解决了搜索过程中出现的局部极小值问题。实验表明该方法能有效地实现多设备医学图象的配准。

In this paper,a contour based rigid registration method of multimodal medical images is pvesented. Based on the semiautomatic search of contour, by using iterative algorithm of closest points to find corresponding points, the cost function is minimized and a best transformation can be obtained, so that the images are registered precisely. Depending on the strategy of choosing initial transformations many a time,the local minimum problem in process of iterative detection have been solved. This method is proved to be effective on registration of multimodal medical images in our experiment.