Multilevel thresholding for image segmentation based on parallel distributed optimization

Abstract

Image segmentation is a platform for many high-level applications. It aims to identify the different homogeneous regions of an image. In our work, we cast the problem as a parallel distributed optimization task and we present a new method for image segmentation by multilevel thresholding based on generalized island model (GIM). This model is characterized by the cooperation of three metaheuristics namely Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Artificial Bee Colony (ABC). The cooperation is ensured by the use of a migration pattern the purpose of which is to avoid getting stuck in local optima. The effectiveness of the proposed method has been assessed and compared to each of the algorithms of the GIM model when used separately and other methods from the literature like Darwinian PSO (DPSO), Fractional-Order Darwinian PSO (FO-DPSO). The obtained results are very promising and show the positive impact of the use of a parallel distributed model. They also show that the proposed method competes and even outperforms the other methods.