INVESTIGATIONS ON ENTROPY BASED THRESHOLD METHODS

Abstract

Entropy is very consistent technique used for image segmentation. In this paper we analyze the performances of three entropic threshold methods against standard Otsu discrimination analysis for threshold selection based image segmentation. Shannon’s entropy, Kapur’s entropy , Fuzzy Entropy are considered to evaluate their performances against Otsu method. The fuzzy S- function is used to calculate the membership degree of each gray value of the image used in computing the entropy of the image. Experimental results demonstrate a quantitative superiority of the Fuzzy entropy over all other methods.