Abstract

Fuzzy C-Means (FCM) algorithm is one of the most popular methods of image segmentation, but it is in essence a technology of searching local optimal solution. The algorithm's initial clustering centers are the stochastic selection which causes it to depend on the selection of the initial cluster centers excessively. FCM algorithm always converges at the local optimum and is sensitive to noise. In order to overcome those defects, this article proposes a PSO based Fuzzy C-Means algorithm. First the overall robustness advantages of the particle swarm algorithm are used to get the cluster centers of image. Then the results are obtained as the initial cluster centers of Fuzzy C-Means algorithm. The experimental results show that new algorithm can converge more quickly than the standard FCM algorithm and suppress the noise effectively.

Paper Details