Image enhancement based on improved pulse coupled neural network

Abstract

Image is the main source of information acquisition and exchange. Image processing technology can realize man-machine dialogue, but it still faces many challenges, such as the lack of image edge information. In order to better realize the neural network image processing, this paper builds two kinds of neural network models based on memristor on the MATLAB platform. Then these two neural network models are used to enhance the image. The image is illustrated by the comparison of the histogram, the enhanced image and some objective criteria through the pulse coupled neural network and the improved neural network. The improved neural network is more efficient than the pulse coupled neural network in processing the edges and internal details of the image, highlighting the outline of the key part of the image, and effectively realizing the image enhancement. The advantages of the neural network with memristor in image processing are illustrated.