FAST IMAGE SEGMENTATION USING C-MEANS BASED

Abstract

In this paper, we propose a fast C-means based training of<br><br>Fuzzy Hopfield neural network and apply it to image<br><br>segmentation. According to the other ways which usually<br><br>take a long time, we define a fast method for image<br><br>segmentation. We present a new objective function, and its<br><br>minimization by Lyapunov energy function which is based<br><br>on two dimensional fuzzy Hopfield neural network. This<br><br>objective function is the same energy function Hopfield<br><br>neural network which is improved, and includes average<br><br>distance between image pixels and cluster centers. In this<br><br>new method, numbers of iterations are less than the other<br><br>methods it means the proposed method has a faster<br><br>convergence rate in comparison with the other ways.<br><br>Therefore, Fuzzy Hopfield neural network method provides<br><br>image segmentation better than the other methods according<br><br>to experimental results.