摘要
如何准确地从CT图像中提取出感兴趣的组织,是医学图像分割中的难点。提出了一种基于模糊连接度的交互式CT图像分割算法:先根据用户指定的感兴趣区域的灰度范围预分割图像,然后用户从结果图像中选择目标和背景种子点,计算出各像素点与两类种子点的模糊连接度,最后根据连接度大小将像素点划分到目标或背景区域。分割过程中,用户可以通过增设或删除目标或背景种子点,修正分割的结果。实验表明,该算法能准确有效地分割出感兴趣区域。
It＇s a difficult task in segmentation of medical images how to extract Region Of Interesting（ROI） from CT.In this paper,an interactive algorithm based on fuzzy connectedness for CT image segmentation is proposed.Firstly,the lower and upper gray values of ROI are specified by users and the algorithm pre-segments the CT image with these values.Secondly,users select seeds from the pre-segrnented image for object and background.The method computes fuzzy connectedness from each pixel to the seeds,and then divides each of them into object area or background by comparing connectedness.Finally,users can improve results by adding or removing object seeds or background seeds.Experiments show that the new algorithm can segment ROI correctly.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第11期216-219,共4页
Computer Engineering and Applications
基金
国家高技术研究发展计划(863)No.2006AA02Z346
湘南学院重点课题(No.Z07009)~~
关键词
模糊连接度
CT图像
区域生长
交互式分割
fuzzy connectedness
CT image
region growing
interactive segmentation
