Image thresholding applications with two dimensional entropy

Abstract

Image thresholding is an important task both for digital image processing applications and for pattern recognition. Image segmentation by thresholding is the simplest technique. In this study, we examine two dimensional entropic thresholding algorithms and we evaluate their superiorities and weaknesses again the one dimensional entropic thresholding algorithms on image thresholding applications. We also examine co-occurrence matrix for forming the two dimensional histogram. The performance evaluations of the examined algorithms are determined based on computer simulations. The two dimensional entropic thresholding schemas give satisfactory results. © 2005 IEEE.