Multilevel Thresholding for Image Segmentation Through an Improved Quantum-Behaved Particle Swarm Algorithm

Multilevel thresholding is one of the most popular image segmentation techniques. Some of these are time-consuming algorithms. In this paper, by preserving the fast convergence rate of particle swarm optimization (PSO), the quantum-behaved PSO employing the cooperative method (CQPSO) is proposed to save computation time and to conquer the curse of dimensionality. Maximization of the measure of separability on the basis of between-classes variance method (often called the OTSU method), which is a popular thresholding technique, is employed to evaluate the performance of the proposed method. The experimental results show that, compared with the existing population-based thresholding methods, the proposed PSO algorithm gets more effective and efficient results. It also shortens the computation time of the traditional OTSU method. Therefore, it can be applied in complex image processing such as automatic target recognition.