Premium
Image thresholding based on gray level‐fuzzy local entropy histogram
Author(s) -
Zheng Xiulian,
Tang Yinggan,
Hu Wenzhao
Publication year - 2018
Publication title -
ieej transactions on electrical and electronic engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.254
H-Index - 30
eISSN - 1931-4981
pISSN - 1931-4973
DOI - 10.1002/tee.22609
Subject(s) - thresholding , histogram , artificial intelligence , balanced histogram thresholding , pattern recognition (psychology) , pixel , fuzzy logic , mathematics , entropy (arrow of time) , image segmentation , membership function , fuzzy set , computer science , segmentation , histogram matching , image (mathematics) , physics , quantum mechanics
The thresholding method utilizes only the gray level information of image but ignores the spatial information between pixels. Thus, it sometimes produces incorrect segmentation results. In this paper, a novel histogram, called gray level‐local fuzzy entropy (GLLFE) histogram, is proposed to incorporate spatial information into the thresholding process. First, the proposed method transfers the pixel's gray level to a fuzzy set through a fuzzy membership function. Second, the local fuzzy entropy of each pixel is calculated and the GLLFE histogram is constructed by combining the local fuzzy entropy and gray level. Finally, a two‐dimensional threshold vector is determined according to the maximum entropy principle. The local fuzzy entropy can not only characterize the spatial correlation but also suppress the noise and enhance the weak edge. Experimental results show that the performance of the proposed method is good. © 2017 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.