自适应加权空间信息的FCM医学图像分割

FCM algorithm with adaptive weighted spatial information for medical image segmentation

提出了改进的mFCM算法,该算法引入自适应加权系数控制邻域像素对中心像素的影响程度,充分利用像素的邻域特性对Chen聚类算法的目标函数进行改进。为了实现快速聚类,该算法的开始使用快速FCM确定初始聚类中心。实验结果表明,相对于标准FCM和FCM_S1算法,改进算法既能快速有效地分割图像,又能提高对噪声的鲁棒性。

The improved algorithm controls the effect of the neighborhood pixels on central pixel by introducing an adaptive weighted coefficient,and modifies the Objective function given in the Chen＇s algorithm.To realize fast clustering,the beginning of the algorithm uses fast FCM. Experiment results show that the improved algorithm is more efficient and more robust to noise than the standard FCM and FCM_ S1.