基于改进的模糊C均值聚类图像分割新算法

A New Algorithm for Image Segmentation Based on Modified Fuzzy C-Means Clustering

模糊C均值(FCM)聚类算法广泛用于图像的自动分割,但是传统的FCM算法没有考虑像素的空间信息,因而对噪声十分敏感。为了克服上述问题,提出了一种新的基于改进的FCM图像分割算法。该方法将空间的信息融入到标准的FCM算法中,通过引入表征邻域像素对中心像素作用的先验概率来重新确定当前像素的模糊隶属度值,该概率在算法执行过程中根据模糊隶属度值自动地予以确定。算法中使用基于统计直方图的快速FCM算法进行初始化,收敛速度大大提高。人造图像和实际图像的实验结果表明该方法的有效性和对噪声具有较强的鲁棒性。

Fuzzy c-means（FOM） clustering algorithm has been widely used in automated image segmentation. However,the conventional FOM algorithm is noise sensitive because of not taking into account the spatial informations. To overcome the above problem,a novel modified FOM algorithm for image segmentation is presented in this paper. The algorithm is formulated by incorporating the spatial neighborhood into the standard FCM clustering algorithm. A prior probability is given to indicate the spatial influence of the neighboring pixels on the centre pixel,which is automatically determined in the implementation of the algorithm by the fuzzy memberships of the neighboring pixels. The new fuzzy membership of the current pixel is then recounted with the obtained probability. The algorithm is initialized by a statistical histogram based on FOM algorithm,which can speed up the convergence of the algorithm. The new algorithm is applied to synthetic and real images and is shown to be effective and more robust to noise and other artifacts than the standard FOM algorithm.