Research on image filter template based on cellular neural network

Abstract

Image filtering is an indispensable task in image processing. Although the existing image filtering methods improve the effect of image filtering, most of them ignore the running time of the algorithm. In order to solve this problem, this paper applies the cellular neural network (CNN) to achieve image filtering, and designs three image filter templates: mean filter, median filter and Gaussian filter. Compared with the traditional mean filter, median filter, Gaussian filter, iterative mean filter (IMF), adaptive median filter (AMF) and Gaussian curvature filter (GCF), the simulation results show that the proposed method achieves the same level as the traditional methods in mean square error (MSE), peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM). The running time of the mean filter and Gaussian filter based on CNN are less than that of traditional mean filter and Gaussian filter. In addition, the three filter templates proposed in this paper are better than IMF, AMF and GCF in dealing with Gaussian noise and multiplicative noise. Running time of the algorithm is also significantly lower than that of IMF, AMF and GCF. The mean filter, median filter and Gaussian filter are implemented by using the theory of cellular neural network.