Elephant Herding Optimization for Multi-Level Image Thresholding

Elephant Herding Optimization for Multi-Level Image Thresholding

Falguni Chakraborty (Nit Durgapur, Durgapur India), Provas Kumar Roy (Department of Electrical Engineering, Kalyani Government Engineering College, Kalyani, West Bengal, India) and Debashis Nandi (NIT Durgapur, Durgapur, India)
Copyright: © 2020 |Pages: 27
DOI: 10.4018/IJAMC.2020100104
OnDemand PDF Download:
No Current Special Offers


Multilevel thresholding plays a significant role in the arena of image segmentation. The main issue of multilevel image thresholding is to select the optimal combination of threshold value at different level. However, this problem has become challenging with the higher number of levels, because computational complexity is increased exponentially as the increase of number of threshold. To address this problem, this paper has proposed elephant herding optimization (EHO) based multilevel image thresholding technique for image segmentation. The EHO method has been inspired by the herding behaviour of elephant group in nature. Two well-known objective functions such as ‘Kapur's entropy' and ‘between-class variance method' have been used to determine the optimized threshold values for segmentation of different objects from an image. The performance of the proposed algorithm has been verified using a set of different test images taken from a well-known benchmark dataset named Berkeley Segmentation Dataset (BSDS). For comparative analysis, the results have been compared with three popular algorithms, e.g. cuckoo search (CS), artificial bee colony (ABC) and particle swarm optimization (PSO). It has been observed that the performance of the proposed EHO based image segmentation technique is efficient and promising with respect to the others in terms of the values of optimized thresholds, objective functions, peak signal-to-noise ratio (PSNR), structure similarity index (SSIM) and feature similarity index (FSIM). The algorithm also shows better convergence profile than the other methods discussed.
Article Preview

1. Introduction1

Image segmentation is a technique of separating non-overlapping regions of an image. Since the last decade, researchers are using image segmentation as a preprocessing step of computer vision and image analysis. Therefore, the accuracy of segmentation is a vital factor for higher level image analysis. A lot of methods like thresholding based, edge based, region based, graph-cut based and connectivity-preserving relaxation methods for segmentation of image objects have been evolved till date. However, most of the researchers use thresholding-based segmentation in different applications because of its simplicity, accuracy, and robustness. In bi-level thresholding, image is divided into only two different homogeneous regions by using a single optimized threshold value whereas multi-level thresholding (Yin 1999, Yin 2007, Sathya et. al.2010, Hammouche 2008) divides the image into multiple regions by using the multiple optimized threshold values. Thresholding technique can be categorized into local and global, among the various global thresholding technique. There exist a number of techniques to compute thresholds. Otsu’s between class variance (Otsu, 1979) and Kapur’s entropy (Kapur et al.1985) are two popular of them which are used for image segmentation. The main challenge in any thresholding-based image segmentation is to find out the best set of thresholds for which the partitioning of different regions of an image may provide the best accuracy. Since the performance of classical optimization techniques broadly depend on the nature of the problems and type of objective functions, their efficiency largely depends on the solution space and the number of variables.

To overcome these drawbacks nature-inspired metaheuristic algorithms have become a research interest of many researchers in the recent days. For example, genetic algorithm (GA) (Yin, 1999; Hammouche, 2008), improved GA (Zhang et al., 2014), biography based optimization (BBO) (Simon, 2008), particle swarm optimization (PSO) (Yin, 2007; Kennedy et al., 1995; Shi, 2001), artificial bee colony (ABC) (Akay, 2013), modified ABC (Bhandari et al., 2015), differential evolution (DE) (Gandomi et al., 2012), bacterial foraging optimization (BFO) (Sathya et al., 2010), ant colony optimization(ACO) (Tao et al., 2007), cuckoo search (CS) (Agrawal et al., 2013), honey bee mating optimization (HBMO) (Horng, 2010), social spider optimization (SSO) (Ouadfel et al., 2016), flower pollination (Ouadfel et al., 2016), BAT (Yang, 2010) algorithm etc. have been developed and successfully applied to solve many optimization problems in different fields. But, in line with the no-free-lunch theorem (Wolpert et al., 1997), a metaheuristic algorithm may not always be able to produce the best result in solving all types of optimization problems. Therefore, we need to continually search a better metaheuristic algorithm which can be able to provide better results in solving a particular problem.

Complete Article List

Search this Journal:
Volume 13: 4 Issues (2022): 2 Released, 2 Forthcoming
Volume 12: 4 Issues (2021)
Volume 11: 4 Issues (2020)
Volume 10: 4 Issues (2019)
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing