The shadows in liver are likely to be cancer tissue, so the liver cancers can be detected by detecting the shadows first. The shadows and normal liver tissue are hard to divide for the densities of them are close to each other. An algorithm is proposed to calculate the signed distance function (SDF) in Chan- Vese model. This algorithm makes the initial curve close to the division edge at the beginning, largely decreases the repeating times and improves the speed of target detection of Chan-Vese model. Finally, this algorithm is applied to the accurate detection of shadows in liver CT images and divides the shadows from normal liver tissue successfully.%肝部阴影是肝癌的重要疑似对象,因此检测肝癌可以先将肝部阴影检测出来.但肝部阴影与正常肝组织密度接近,很难分割.为此,提出一种CV分割方法的初始距离函数的设置方法,该方法使初始曲线一开始就接近图像的分割边缘,极大地减少了CV方法的迭代次数,加快了CV方法目标提取的速度.最后将该方法应用于肝部阴影的精确检测,成功地分割了图像中的肝部组织与其中的阴影部分.
展开▼
