Fuzzy Clustering Segmentation Algorithm Research on Sports Image

Abstract

In this paper, linking with the basic principle of FCM (Fuzzy c-means clustering) algorithm, on the basis of theory research, a method of the cluster analysis of FCM is proposed. Firstly, the approximate optimal solution obtained by the improved FCM algorithm is taken as the original value of the FCM algorithm, then carrying on the local search to obtain the global optimal solution, the final segmentation result is achieved at last. The experiment results prove that in the view of the flame image segmentation, this method shows the good clustering performance and fast convergence rate, and has the widespread serviceability, so it is the practical method in image segmentation.