基于边缘信息C_V模型的医学图像分割方法  

Boundary Information Based C_V Model Method for Medical Image Segmentation

作　　者：秦绪佳[1] 杜轶诚[1] 张素琼[1] 王卫红[1] 韩军 

机构地区：[1]浙江工业大学计算机科学与技术学院,浙江杭州310032 [2]大连现代高技术公司数字核医疗部,辽宁大连116025

出　　处：《小型微型计算机系统》2011年第5期972-977,共6页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61075118;60873033)资助;国家科技支撑计划项目(2007BAH11B02)资助;浙江省自然科学基金项目(Y1100880)资助;浙江省科技计划项目(2009C31106)资助

摘　　要：图像分割是医学处理中的重要研究内容之一,提出一种基于边缘信息的改进的C_V模型的医学图像分割方法.在模型中增加了表征边界特征的项,利用图像的边界信息与区域信息为分割服务,克服了传统C_V模型不能利用图像的梯度信息的不足.并对C_V模型的区域信息项进行了改造,改变了传统C_V模型中均值取值的定义,提高了对灰度层次丰富的图像分割能力.增加了距离函数惩罚项,将距离函数重新初始化的过程并入整个水平集框架模型中,极大地提高了曲线演化与分割速度.实验表明该模型是有效的医学图像分割方法.The image segmentation is one of the key problems in medical image processing.An improved C_V（Chan Vese） model for medical image segmentation based on boundary information is proposed.Firstly,a term of boundary information is added into the model,incorporating region and boundary information for segmentation.It solves the problem that the traditional C_V method can not use the gradient information.Secondly,the region information term and the mean value definition of the whole image in the traditional C_V model have been changed.It increases the segmentation ability of rich levels gray image.Finally,to overcome the re-initialization,a penalty term of distance function is added into the model,the progress of re-initialization is combined into the framework model.It can speed up the curve evolution and the segmentation.The experiments show that the model is an effective method for medical image segmentation.

关 键 词：水平集方法 C_V模型 重新初始化 医学图像分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]