改进的区域生长算法及其在肾实质自动分割中的应用  

An Improved Region Growing Algorithm and Its Applications in Kidney Segmentation

作　　者：高岩[1] 王博亮[1] 

机构地区：[1]厦门大学信息科学与技术学院,福建厦门361005

出　　处：《厦门大学学报（自然科学版）》2012年第4期701-703,共3页Journal of Xiamen University：Natural Science

基　　金：国家自然科学基金项目(30770561;61071151;61102137)

摘　　要：提出了一种肾脏CT图像自动分割方法,将像素点的局部统计特征和像素点的空间位置信息结合起来,以此定义了像素之间的邻域相似性指标,并根据领域相似性指标自动选取种子点、种子的生长准则及终止准则,该方法克服了传统区域生长算法需手动确定种子点和生长顺序固定等缺点,最后通过MICCAI(medical image computing and computer assited intervention)的5个评价指标对分割结果做出客观评价,结果表明,该算法具有较好的分割效果.An improved region growing algorithm for automatic segmentation kidney from abdominal CT image is proposed. Com-pared to original region growing method,this method automatically selected initial seed pixels and is robust to the order of region growing. Firstly,computing the neighbor similarity factor （NSF） based on local histogram of each pixel and spatial information of lo cal pixels. Then, building the criteria of initial seed-pixels, region growth and region growing termination based on NSF. Finally, MIC CAI metrics are adopted to measure the segmentation accuracy. Experimental results demonstrated the performance of method.

关 键 词：生长算法 自动分割 肾脏 应用 评价指标 IMAGE 像素点 种子点 

分 类 号：R318[医药卫生—生物医学工程] TP391[医药卫生—基础医学]