基于最大流的交互式目标提取算法  

Iteractive object extraction based on Max-flow

作　　者：于丹[1] 汤井田[1] 徐大宏[1] 罗东礼[1] 

机构地区：[1]中南大学信息物理工程学院生物医学工程研究所,长沙410083

出　　处：《计算机工程与应用》2007年第5期246-248,共3页Computer Engineering and Applications

基　　金：国家"十五"科技公关计划项目(the Key Technologies R&D Program of the "Tenth Five-Year-Plan" of China under Grant No.2001BA609A-5)

摘　　要：医学图像处理中,目标提取在准确区分组织结构中起了重要的作用,论文介绍了基于最大流的交互式的目标提取算法,该方法把基于区域和基于边界方法结合起来,并用分水岭算法对图像做预分割,在其所分的区域上计算能量函数,然后最小化能量函数求最优边界。笔者结合自己研究的课题,针对目标标记步骤中聚类种子F和#色彩值的k-means算法进行改进,并把改进算法应用到人体器官切片目标分割中。Object extraction plays an important role in distinguishing organize framework accurately in medical image process.This text has introduced an interactive object extraction on the basis of max-flow,this method combines the basis of area with the basis of border method,and using watershed algorithm for image pre-segmentation,then calculating energy function on the segmented regions from it.The author uses one＇s own subject for research,applies this kind of algorithm to the human organ seg- mentation and improves in the k-means method by which the colors in seeds F and β are clustered in object marking.

关 键 词：目标提取 最大流 图切割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]