Multi-threshold Image Segmentation Based on Firefly Algorithm

Abstract

Multi-threshold image segmentation have achieved good results, but multi-threshold searching can be very time consuming, in order to improve the efficiency of multi-threshold searching, in this paper, the Firefly Algorithm is applied to solve multi-threshold searching. With maximum entropy as the optimization objective function, more information of images can be reserved, then using the Firefly Algorithm's heuristic searching, the optimal thresholds of image segmentation can be found. Experimental results show that compared to Genetic Algorithm, our proposed Firefly Algorithm has a higher convergence speed and produces a better segmentation effect when applied to image segmentation.