Affordable Access

Access to the full text

Multi-resolution gray-level image enhancement using particle swarm optimization

Authors
  • Nickfarjam, Ali Mohammad1
  • Ebrahimpour-Komleh, Hossein1
  • 1 University of Kashan, Faculty of Electrical and Computer Engineering, Computer Engineering Department, Kashan, Iran , Kashan (Iran)
Type
Published Article
Journal
Applied Intelligence
Publisher
Springer US
Publication Date
May 15, 2017
Volume
47
Issue
4
Pages
1132–1143
Identifiers
DOI: 10.1007/s10489-017-0931-2
Source
Springer Nature
Keywords
License
Yellow

Abstract

This paper presents a multi-resolution method for gray-level image enhancement using Particle Swarm Optimization (PSO). The enhancement optimization procedure is a non-linear problem with various constraints. The proposed image enhancement algorithm (MGE-PSO) generates a whole pyramid of differently sized image in order to utilize more information for improvement process. In fact, MGE-PSO employs the ability of image pyramid to determine informative parts of an image for visual perception. When an image is downscaled, area of homogeneous regions is decreased and informative pixels of input image can be selected easier. The PSO uses averaged variance value of all pixels included in the informative and non-informative classes of each level in image pyramid to move through search space for finding the best intensity values of pixels to transfer maximum visual perception. Experimental results on Berkeley dataset demonstrate the superiority of the proposed MGE-PSO to other methods. Beside, detailed analysis of selection criterion used in PSO are available.

Report this publication

Statistics

Seen <100 times