摘要
医学图像配准是医学图像融合的前提,是目前医学图像处理中的热点,具有重要的临床诊断和治疗价值。依据7个标准对医学图像配准方法进行了分类,然后从3个方面综述了目前的一些主要的医学图像配准方法,如矩和主轴法、互信息法和相关法等。由于医学图像配准是一项比较复杂和困难的课题,尽管目前已提出许多算法,但并没有一种配准方法能在各个方面都达到理想要求,因此不得不在精确度、速度、自动化程度等方面加以取舍。随着计算机技术和医学成像技术的发展,医学图像的配准技术也一定会得到快速发展。
Medical image registration, which is the precondition of medical image fusion, is a highlight of present medical image processing, and has important values in clinical diagnosis and treatment. This paper firstly presents a classification of medical image registration methods in terms of seven standards, then summarizes some major registration methods from three aspects, such as moment and principal axes method, mutual information, and correlation method etc. Although there are many algorithms at present, no registration method could come up with ideal requirements in every aspect because of its difficulty and complexity. Consequently registration method has to make its choice between accuracy, velocity, automation extent and other aspects. However, with the development of computer technology and medical imaging, the registration technology of medical image is bound to develop quickly.
出处
《重庆大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2003年第8期114-118,共5页
Journal of Chongqing University
关键词
图像配准
图像融合
矩和主轴法
互信息
相关法
医学图像
image registration
image fusion
moment and principal axes method
mutual information
correlation method
