基于蚁群行为及模糊聚类的乳腺钙化点边缘提取

针对乳腺X光图像中钙化点形态各异、周围背景复杂及其与背景对比度不同造成的钙化点识别难度大的问题,模仿蚁群行为对乳腺X光图像中各像素点按梯度值进行聚类,输出聚类中心;以这些聚类中心作为模糊C均值法(FCM)的初始聚类中心,经迭代得到新的聚类,再经过适当的类间合并,从而将钙化点边缘的像素提取出来.实验结果表明,通过选择适当的蚂蚁聚类及模糊C均值法参数,该方法可有效提取乳腺钙化点边缘.

Edge detection of mammary calcifications based on behavior of ant colony and fuzzy Clustering

Detection of mammary calcifications is challenging because of their different shape,complicated background and different contrast between calcifications and their background.Clustering of pixels mammographic of image is implemented by ant based clusterg according to gray grads and initial cluster centers are obtained.New clustering result is obtained through fuzzy c-means algorithm by using the above cluster centers as initial values.Edges of mammary calcifications are extracted through appropriate combinat...