结合图像熵的C-V模型分割算法  

Algorithm of C-V model combined with image entropy

作　　者：叶丙刚[1] 姜涛[1] 罗朝辉[1] 吴效明[1] 

机构地区：[1]华南理工大学生物科学与工程学院,广州510006

出　　处：《计算机应用》2009年第12期3320-3321,3339,共3页journal of Computer Applications

基　　金：国家自然科学基金资助项目(60772037);江西省教育厅科技计划项目(GJJ09344)

摘　　要：C-V模型是水平集分割方法中的一种经典模型,但存在自适应能力不强的问题。医学图像分割对象的特点是具有各种复杂的拓扑结构及变化。根据医学分割算法的自适应要求,在水平集图像分割方法C-V模型算法的研究基础上,引入图像信息熵算法,通过计算演化曲线内外的图像熵,解决曲线演化过程中迭代参数设定问题,达到增强C-V模型分割算法的自适应能力。实验结果表明,结合图像熵的C-V模型算法面对不同的对象具有良好的自适应性。C-V model, as a classical one in the level set segmentation methods, also has a weak self-adaptive capability. The medical segmentation objects have diversified complex topologic structures and changes. According to the adaptive requirements of medical segmentation algorithm, the article introduced image information entropy algorithm based on the research on C-V model algorithm of the level set picture segmentation methods, solved the iterated parameters setting issue in the curve evolvement, and enhanced the adaptive capability of C-V model segmentation algorithm. The experiments show the algorithm, C-V model combined with image entropy, could have good adaptability to different objects.

关 键 词：水平集 C—V模型 图像信息熵 迭代参数 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]