Boosted Pathfinder Algorithm and Firefly Algorithm for Multi-level Thresholding Image Segmentation Using Fuzzy Entropy

Abstract

Image processing is very important for many applications in computer vision and digital intelligence society applications. This study presents an alternative multilevel thresholding algorithm for image segmentation. The developed approach aims to improve a metaheuristic (MH) technique called pathfinder (PF) algorithm using firefly algorithm (FA) as a local operator. In the proposed algorithm, called PF-FA, the FA is employed to boost the exploitation ability of the traditional PF algorithm, which is used as a local operator. To test the performance of the developed PF-FA approach, we use a number of gray-scale images. Moreover, the developed PF-FA method is compared to some well-known  MH techniques that are generally employed for multilevel thresholding at different threshold levels. The experimental results revealed that the PF-FA has a competitive performance, and the applications of FA have a significant impact on the traditional PF algorithm.