摘要
为自动获取肝脏纤维化定量分析指数——计算机形态学评分(CM评分),提出一种肝脏病理切片图像的分割方法.该方法首先利用窗口纹理粗糙度图像分割背景,然后利用Fisher准则函数选取最优颜色特征空间分割纤维,最后由分割结果求得纤维与组织的面积比(CM评分).实验结果表明,用该方法自动获取CM评分可大大地提高肝脏纤维化临床诊断的客观性和效率.
To automatically obtain the quantitative index of the liver tissue fibrosis, or computer morphometry (CM) score, an approach is proposed to segment liver pathological sections. By the approach, background regions are first removed based on the window texture coarseness image, and a new way based on Fisher criterion to optimize the color-character space and segment fibrosis areas is implemented. Then, we could calculate the fibrosis-tissue area ratio (CM score) based on the segmentation results. The experimental results indicate that CM scores acquired by using this approach greatly improved the objectivity and the efficiency of the liver fibrosis diagnosis in the traditional clinical methods.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2007年第6期775-780,共6页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(60542003)
国家"八六三"高技术研究发展计划(2006AA02Z347)
北京市科技计划项目(H020920020290)
关键词
肝脏纤维
图像分割
窗口纹理粗糙度图像
FISHER线性判别
liver tissue fibrosis
image segmentation
window texture coarseness image
Fisher linear discriminant analysis