说明：  图像配准在医学图像处理领域是一项重要的技术，对临床诊断和治疗起着越来越重要的作用。尽管对 这方面的研究已经展开多年，但目前的主要方法仍然存在不足之处，急需改进，以便更好的应用于临床实践。 本文主要针对现在流行的基于最大互信息量的配准方法展开讨论和研究。在此基础上提出了相同重叠区域下的配准框架，在此框架下，将一些统计相似性配准算法统一为基于最小条件熵的图像配准算法。通常称为归一化互信息配准的方法。
(Image registration in medical image processing is an important technique for clinical diagnosis and treatment plays an increasingly important role. Although research in this area has started for many years, but there are still the main method of the inadequacies of the urgent need to improve, in order to better clinical practice. In this paper, for now the most popular mutual information based registration method to discuss and research. On this basis, the same overlap region proposed registration under the framework, in this framework, some of the statistical similarity matching algorithm uniform minimum conditional entropy-based image registration algorithm. Often referred to as normalized mutual information registration method.)