Performance Analysis of Image Restoration Filters for Textual Images

Abstract

In the modern era, images are one of the most important media used to store, transmit and analyze the information within it. True images often get degraded during the acquisition process due to environmental disturbances. Image Restoration techniques can be applied to restore such images to extract the information contained within the original data. Over the past years, several restoration techniques have been developed. In this paper, analysis of some of those techniques which are Wiener Filter, Richardson-Lucy Algorithm, Inverse Filter and Median Filter are performed for blur and noisy blur images. These methods have been utilized to restore the deteriorated images. Their performances are compared using factors like Power Signal to Noise Ratio (PSNR), Structural Similarity Index (SSI), and Mean Squared Error (MSE). These algorithms have been implemented on MATLAB(2020a) and the results show that  Wiener Filter gives the best results for even blurred and noisy images.