Optimal multilevel thresholding using the modified adaptive particle swarm optimization

Abstract

The conventional multilevel thresholding methods are computational expensive since they exhaustively search the optimal thresholds to optimize the objective functions. In this paper, the modified adaptive particle swarm optimization (MAPSO) algorithm is proposed to overcome this drawback. The dynamic population (DP) strategy of the proposed algorithm enables the population size variable with the evolutionary state at run time. With the help of DP strategy, the population size can increase when the swarm converges and decrease when the swarm disperses. The MAPSO algorithm is used to find the optimal thresholds by maximizing the Otsu's objective function. The performance of the proposed algorithm has been validated on eight standard test images. The experimental results of 50 independent runs illustrate the best solution quality and stability of the MAPSO when compared with three other PSO algorithms.