基于形状匹配的快速图像配准  

Fast Image Registration Based on Shape Matching

作　　者：张素[1] 苏和[1] 陆雪松[1] 阳维[1] 

机构地区：[1]上海交通大学生物医学仪器研究所,上海200240

出　　处：《天津大学学报》2008年第4期433-438,共6页Journal of Tianjin University(Science and Technology)

基　　金：国家重大基础研究计划(973)资助项目(2003CB716103);上海市科委重点攻关项目(05DZ19509)

摘　　要：为了发挥基于特征的图像配准方法和基于像素的配准方法的优点,提出了一种将特征匹配和最大互信息法相结合的配准方案:先提取参考图像和浮动图像中的目标形状进行匹配,利用匹配结果求出互信息搜索算法的初始值再进行搜索.实验表明,该方法计算量小、速度快且精度高,可避免参数搜索陷于局部极值,有效地提高了配准的速度和精度.To take advantage of feature-based image registration and pixel-based image registration, a new method was proposed which combined the maximum mutual information optimization with shape matching. The target shapes were firstly extracted from reference image and floating image, and then matching was performed. The initial value of the mutual information optimization was calculated using the matching results to continue searching. Experimental results indicate that the proposed algorithm has low computational complexity, fast speed, and high accuracy,moreover it can avoid the convergence at the local extremum when performing parameters optimization, which improves the speed and accuracy of the registration effectively.

关 键 词：形状匹配 互信息 图像配准 

分 类 号：R319[医药卫生—基础医学]