一种基于水平集的医学图像分割算法  

A Segmentation Algorithm for Medical Images Based on Level Sets

作　　者：李玥[1] 唐庆玉[1] 

机构地区：[1]清华大学生物医学工程系,北京100084

出　　处：《北京生物医学工程》2005年第5期370-372,共3页Beijing Biomedical Engineering

摘　　要：对Chan和Vese提出的活动轮廓线图像分割算法进行改进,用边界统计特性代替Chan-Vese方法中的区域统计特性。在图像分割的每次迭代中,首先根据当前曲线找到其边界统计特性,然后代入演化方程以使曲线运动到目标边界上,并用水平集方法实现。实验结果表明该算法在器官轮廓分割中是有效的。The active contour method proposed by Chan and Vese is improved. The boundary statistics is used instead of the region statistics in Chan-Vese＇ s method. In every iteration of image segmentation, firstly the corresponding boundary, statistics of the current curve is found, then it is driven in evolutivc equation,finaly the curveis moved to capture the object＇ s boundary under the level sets framework. Experimental result shows the effectiveness of the proposed method.

关 键 词：活动轮廓线 图像分割 水平集 

分 类 号：R318.04[医药卫生—生物医学工程]