摘要： 提出一种基于Rényi熵最小生成图和multi-quadric径向基函数的医学图像自动弹性配准方法。应用图像金字塔的思想，对图像分层分块，将Rényi熵最小生成图作为相似性测度对子块进行配准，在对应的子块中选取对应的标记点，用multi-quadric径向基函数插值这些标记点，从而实现医学图像弹性配准。实验结果表明，该方法配准速度较快，精度较高，是一种有效的自动弹性配准方法。

关键词: Rényi熵, 熵最小生成图, 医学图像弹性配准, multi-quadric方法

Abstract: This paper presents an automatic elastic registration of medical images based on Rényi entropic minimum spanning graph and multi-quadric radial basis function. Images are layered and blocked based on image pyramid process and its sub-image blocks are registered using Rényi entropic minimum spanning graph’s similarity measure. Landmark points in corresponding sub-image blocks are extracted. Those landmark points are interpolated via multi-quadric radial basis function so as to achieve elastic registration of medical image. Experimental results indicate that the method is a fast, accurate and effective elastic registration of image.

Key words: Rényi entropy, entropic minimum spanning graph, elastic registration of medical images, multi-quadric method

中图分类号: 

• TP391.41