Robust RGB Image Thresholding with Shannon’s Entropy and Jaya Algorithm

Abstract

Image thresholding is a common pre-processing procedure implemented in various domains to improve the picture quality. In this work, thresholding of benchmark RGB-scale picture is implemented with the Jaya Algorithm (JA) and Shannon’s Entropy (SE). The investigational work is executed using Matlab7 software on RGB pictures of varied sizes. This work examines the original and the noise stained pictures based on the chosen threshold. In order to assess the performance of JA+SE, Picture-Quality-Measures (PQM) are computed by comparing the original and thresholded pictures. The average outcome confirms that, JA+SE provide better PQM values in both the normal and noise corrupted cases. Hence, this study confirms that, pre-processing with SE helps to attain better values of the error and PSNR values for the considered images of different size and quality.