一种重叠胃腺癌细胞分离线估计算法  

A Estimating Separate Line Algorithm for Overlapping Gastric Adenocarcinoma Cell Image

作　　者：张云[1] 刘建平[1] 

机构地区：[1]国防科技大学机电工程及自动化学院

出　　处：《中国图象图形学报》2007年第11期2058-2061,共4页Journal of Image and Graphics

摘　　要：针对胃腺癌细胞显微图像中经常出现重叠(或粘连)现象,对后续进一步的研究工作产生了很多不良影响,为了准确地分离重叠(或粘连)的胃腺癌细胞,采用了一种数学形态学与改进的B样条活动轮廓模型算法相结合的方法。该算法首先对图像进行二值化处理,然后利用变元腐蚀得到细胞的非线性距离图和背景的非线性距离图,接着对两者进行叠加,即得到一幅新的非线性距离图,最后使B样条曲线准确收敛到目标边界,以便自动分离细胞。实验结果表明,该算法不但能有效地构造分离线,并能估计出重叠(或粘连)胃腺癌细胞的原始形状。Computer-aided medical image processing and analysis is a important application in the field of image processing. The introduction of computer-aided diagnosis system of cancerous tissues is highly helpful to clinical diagnosis and medical studies, especially in the case of lacking specialists. But the technology is difficult to put into practice. For example there are many overlapping cells in the gastric adenocarcinoma image because of the facture of the micrograph of gastric adenocarcinoma. If the cells are not separated accurately, the following research will not go on well. It is important to draw the separate-line and gain the borderline of each cell. So the paper presents a algorithm which integrated mathematical morphology with an improved B-spline active contour model （B-Snake model） in order to segment the overlapping parts in the gastric adenocarcinoma image. The algorithm adopted an improved nonlinear distance image to force B-spline curve close to the objective boundary automatically. The experimental result of shows that our algorithm can construct the separate-line exactly and draw the original figures of cells effectively.

关 键 词：数学形态学 B-SNAKE 分离线 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]