摘要
相似性测度作为一种准则用来评价参考图和变换后图像之间的匹配效果,是图像配准方法中一个关键的步骤,它直接影响配准效果的好坏.在分析图像联合直方图的基础上提出一种基于联合直方图的相似性测度,该方法提高配准的计算速度而不影响配准效果.所使用的相似性测度应用于肝脏CT增强扫描多相期刚性图像配准过程,实验结果表明,其速度优于常用的最大互信息法,且不影响配准结果.
As a principle to evaluate the matching effect between target image and transformed image,the similarity metric,which is a critical step in the image registration,will affect registration results directly.In the analysis of the joint distribution histogram of images,a similarity metric based on joint distribution histogram is proposed for speeding up the computation without impacting registration results.Applied to live CT enhanced scanning multiphase images registration,as experimental results show,the proposed method computes faster than the commonly-used maximized mutual information method and does not affect matching effects the registration results are identical.
出处
《厦门大学学报（自然科学版）》
CAS
CSCD
北大核心
2015年第3期397-403,共7页
Journal of Xiamen University：Natural Science
基金
国家自然科学基金(61102137
61271336
61327001)
关键词
图像配准
相似性测度
联合直方图
image registration
similarity metric
joint distribution histogram
