Modified Fuzzy Hopfield Neural Network Using for MRI Image segmentation

In this paper, a new method Based on fuzzy Hopfield neural network for segmentation of MRI brain images is proposed. In MRI Images, the noise of imaging process can cause error in the conventional intensity base classification methods. In our algorithm, by optimizing the energy function fuzzy Hopfield neural network, we considered the effect of the neighbors of a pixel in classification of that pixel, and therefore we modify the structure of this network against the effect of noise problem. In other word in our method the labeling pixel has been influenced by the labels in the immediate neighborhoods. The results obtained from the proposed algorithm show that this method has an acceptable accuracy for segmentation of the brain´s images.