Image thresholding via a modified fuzzy c-means algorithm

Abstract

In this paper, a modified fuzzy c-means (FCM) algorithm named weighted fuzzy c-means (WFCM) algorithm for image thresholding is presented. The algorithm is developed by incorporating the spatial neighborhood information into the standard FCM clustering algorithm. The weight indicates the spatial influence of the neighboring pixels on the centre pixel, which is derived from the k-nearest neighbor (k-NN) algorithm and is modified in two aspects so as to improve its property in the WFCM algorithm. To speed up the algorithm, the iteration in FCM algorithm is carried out with the statistical gray level histogram of image instead of the conventional whole data of image. The performance of the algorithm is compared with those of an existing fuzzy thresholding algorithm and widely applied between variance and entropy methods. Experimental results on both synthetic and real images are given to demonstrate the proposed algorithm is effective and efficient. In addition, due to the neighborhood model, our method is more tolerant to noise. © Springer-Verlag 2004.