2-D Maximum-Entropy Thresholding Image Segmentation Method Based on Second-Order Oscillating PSO

The image segmentation based on 2D histogram which considered gray information as well as spatial neighbor information between pixels of an image, was an efficient image segmentation method yet requires a large amount of computing time. By convergence analysis and simulation, the optimal value to study factor of second-order oscillating particle swarm optimization (SOPSO) algorithm was proposed firstly and it was different from standard PSO. The algorithm was applied to 2D maximum-entropy thresholding image segmentation. The simulation results showed that the algorithm could find the optimum 2D thresholding of an image rapidly and stably and the segmentation results of the Lena picture was ideal.