A Comparative Study on Multilevel Thresholding Using Meta-Heuristic Algorithm

Abstract

Current research work is designing the biological visual system which can emulate the human visual system. Image segmentation is one of the important initial steps in this area. There are different approaches to perform segmentation. One of the well-known techniques in image segmentation to separate objects from the background is Image thresholding. Segmentation using multiple thresholds is treated as an optimization problem in most of the cases. This can be done by maximizing or minimizing a given objective function. This paper presents a comparison of seven well known meta-heuristic techniques to obtain optimal threshold for multilevel thresholding problem: wind driven optimization, grey wolf optimization, firefly algorithm, whale optimization, crow optimization algorithm, and grasshopper optimization. Experimental results present the quantitative and qualitative measures of the different algorithms on multi-level thresholding problem with advantages and drawbacks.