Artificial Neural Networks Evaluation as an Image Denoising Tool
Yazeed A. Al-Sbou · 2012
Abstract: Image denoising is a challenging task in the digital image processing research and application. This makes it imperative to find a robust method to comply that task. In this paper, a detailed performance evaluation of using the neural networks as a noise reduction tool is presented. The proposed approach includes using both mean and median statistical functions for calculating the output pixels of the training pattern of the neural network. This uses part of the degraded image pixel to generate the system training pattern. Different test images, noise levels and neighborhoods sizes are used. Based on using samples of degraded pixel neighborhoods as inputs, the output of proposed approach provided a good image denoising performance, which exhibited promising qualitative and quantitative results of the degraded noisy images in terms of PSNR, MSE and visual tests. Key words: Image denoising Neural networks Pixel neighborhoods Noise variance PSNR MSE