一种改进的顾及像素空间信息的FCM聚类算法

Novel modified fuzzy c-means clustering algorithm considering pixel spatial information

标准的FCM算法对噪声比较敏感,主要是因为该算法没有考虑像素间的空间信息。为了克服这个不足,本文基于自适应加权均值滤波图像提出了一种用于图像分割的FCM改进算法。该算法通过修改Ahmed聚类算法中的目标函数实现。利用该改进算法进行合成图像和真实图像的实验结果表明,相对于标准的FCM聚类算法和由Ahmed改进的算法,本文中提出的改进算法对于噪声更具有鲁棒性。

Standard FCM algorithm is noise sensitive because of not taking into account the spatial information. To overcome the above problem, this paper proposes a modified FCM algorithm for image segmentation based on adaptive weighted averaging image. The presented algorithm was realized via modifying the objective function given in the Ahmed＇s algorithm. Experimental results on both synthesized and realistic images show that the presented algorithm performs more robust to noise than the standard FCM algorithm and another FCM algorithm （proposed by Ahmed） do.