Image Segmentation Based on 2-D Fisher and Chaos Particle Swarm Optimization Algorithm

Abstract

Aimed at solving the deficiencies of having great computation and local optimum, a chaos particle swarm optimization algorithm which has the properties of ergodicity, randomicity and regularity is proposed to solve the optimization problems cited from pattern recognition, and a new method of image segmentation using two-dimensional histogram based on Fisher criterion function is put forward. According to the experiment, it verifies that the method of chaos particle swarm optimization can improve the efficiency and accuracy of 2-D fisher method based on traditional PSO.