摘要
定义了基于形状特征点的互信息计算公式 ,提出区域增长结合动态聚类算法的形状特征点提取方法 .在使形状特征点互信息最大化完成医学图像配准的过程中 ,引入人机交互 ,缩短了优化过程 ,避免了局部极值 .提出的配准策略具备临床实用性 。
Mutual inform ation representing the anatom ic features of medical im age is com puted with feature points. The feature points are extracted from images using region growing and dynamic clustering. During maxim ization of the mutual information to accomplish image registration,man- machine interaction is used to facilitate this procedure and a ransack method is also adopted to avoid local extrema.Prelim inary results on two- dimensional robust alignm ent of MRI- MRI images and MRI- CT images are presented.The registration strategy presented here m ight be suitable to clinical applications,especially for images without gray level information.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2002年第7期654-658,共5页
Journal of Computer-Aided Design & Computer Graphics
基金
北京市自然科学基金 (3 982 0 0 2
3 0 2 2 0 0 4)
卫生部科学研究基金
国家"八六三"高技术研究发展计划 (2 0 0 1AA2 3 10 3 )资助
