摘要
提出一种基于区域生长的CT序列图像的分割算法。在第一张待分割目标区域中选取一个种子点,利用四领域的生长规则对种子点进行区域生长,得到一组点集,将这个点集投影到下一张CT图像中,得到一组新的点集,再提取该点集的轮廓,最后对该轮廓上的点进行四领域的区域生长,分割出最终的目标区域。实验结果表明,该分割算法不仅适用于简单的图像分割问题,而且对于背景复杂、光照不均匀的图像也能取得较好的分割效果。
This paper puts forward a algorithm for segmenting CT sequence image based on region growing. Firstly, selectes a seed from the aim region to be segmented in the first CT image, makes the seed grow based on four neighbors method and get a aggregate of points. Then let the aggregate project to the next CT image and gain a new aggregate of points .Finally distill the contour from the new aggregate and make the points of the contour grow based on the same method as far as get the final aim region. With this method,the target can be extracted without any pixel of the background.
出处
《计算机与数字工程》
2007年第5期1-2,共2页
Computer & Digital Engineering
基金
广东省教育厅自然科学基金(编号:ZL03090014)资助
关键词
CT序列
区域生长
种子点
轮廓
生长规则
CT sequence,region growing,seed point,contour,growing strategy
