Image Segmentation Using Electromagnetic Field Optimization (EFO) in E-Commerce Applications

Image Segmentation Using Electromagnetic Field Optimization (EFO) in E-Commerce Applications

Pankaj Upadhyay (National Institute of Technology Kurukshetra, Kurukshetra, India) and Jitender Kumar Chhabra (National Institute of Technology Kurukshetra, Kurukshetra, India)
Copyright: © 2019 |Pages: 16
DOI: 10.4018/IJISMD.2019070105
OnDemand PDF Download:
No Current Special Offers


Image recognition plays a vital role in image-based product searches and false logo identification on e-commerce sites. For the efficient recognition of images, image segmentation is a very important and is an essential phase. This article presents a physics-inspired electromagnetic field optimization (EFO)-based image segmentation method which works using an automatic clustering concept. The proposed approach is a physics-inspired population-based metaheuristic that exploits the behavior of electromagnets and results into a faster convergence and a more accurate segmentation of images. EFO maintains a balance of exploration and exploitation using the nature-inspired golden ratio between attraction and repulsion forces and converges fast towards a globally optimal solution. Fixed length real encoding schemes are used to represent particles in the population. The performance of the proposed method is compared with recent state of the art metaheuristic algorithms for image segmentation. The proposed method is applied to the BSDS 500 image data set. The experimental results indicate better performance in terms of accuracy and convergence speed over the compared algorithms.
Article Preview

In recent past, various nature-inspired, bio-inspired and physics inspired optimization algorithms have been proposed for clustering and applied to image segmentation. In the last decade, significant progress in automatic clustering algorithms has been done. In particular, different evolutionary bio-inspired metaheuristics have been proposed to obtain near-optimal solutions for cluster analysis. These metaheuristics mimic the natural phenomenon of evolution, social behavior of swarms, physics-inspired behavior of objects. Bandyopadhyay et al. (2002) proposed a genetic algorithm with a variable length string. They used the Davies-Bouldin index as an objective function to evolve the clusters automatically and demonstrated the improved performance on real and artificial data sets. The proposed algorithm is named as GCUK for unknown k. (Omran et al., 2006) proposed particle swarm optimization-based clustering approach named as DCPSO. Initially, it partitioned data set into larger clusters and then used PSO for optimal cluster evolution for image segmentation. Mukhopadhyay & Maulik (2011) proposed a variable string length genetic fuzzy clustering-based image segmentation for T-1 and T-2 weighted brain MRI images.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 13: 4 Issues (2022): Forthcoming, Available for Pre-Order
Volume 12: 4 Issues (2021): 2 Released, 2 Forthcoming
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