医学图像自动多阈值分割

Multi-threshold value segmentation approach for medical images

针对医学图像的自动多阈值分割问题,采用模糊C-均值(FCM)聚类法找到医学图像的不同组织和背景的聚类中心,再利用二维直方图的方法,找到多阈值分割的各个阈值点进行分割。引用二维直方图的方法可以很好地保留目标的细节信息,更好地抑制噪声。

As a solution to automatic multi-threshold value segmentation of medical image,the Fuzzy C-Means(FCM) algorithm was used to find the different targets and the background cluster center of a medical image.Then the method of two-dimension gray histogram was used to find multi-threshold value break point for segmentation of different targets.The use of two-dimension gray histogram is a better way to retain detailed information and to eliminate noise of image.