Image Segmentation Based On Particle Swarm Algorithm And Its Application

Image segmentation is a key technology of image analysis, understanding and description, the main purpose of image segmentation is extracting the interested regions from an image. So far, image segmentation technology has been widely applied in many areas, such as astronomy, agriculture, industry and medicine. However, the results of traditional image segmentation methods are sometimes difficult to satisfy the needs in some complex problems. As a branch of swarm intelligence algorithm, particle swarm optimization (PSO) algorithm provides a new impetus and direction for the developing of image segmentation.However, premature convergence maybe occurs in later searching stage of PSO algorithm as a result of the loss of population diversity. In this paper, PSO algorithm has been improved according to the problem above, and the improved algorithm is also applied in image segmentation. In order to improve the performance of PSO algorithm, the entire population of basic PSO algorithm is divided into several sub-groups in the iteration process, meanwhile the mutation operation in genetic algorithm is introduced into the algorithm, and the worst sub-group is mutated according to mutation probability. Then,the velocity of particle is updated with some large momentum factor and some simulation experiments are taken on classical test functions to verify the performance of the improved PSO algorithm. At last, the improved algorithm is applied to image segmentation, and the experimental results show that the improved PSO algorithm can enhance the convergence accuracy.It's a hot and difficult issue that how to determine the number of categories of image segmentation automatically. The paper gives a new fuzzy C-means clustering algorithm (FCM) based on the improved PSO algorithm to realize the foregoing goal. The improved FCM algorithm combine the improved PSO algorithm given before with FCM algorithm, and a fuzzy cluster validity index was introduced into the FCM algorithm too. This paper design a visible interface system based on the proposed algorithm for the image segmentation users, and a blood cell image is processed by the designed system. The experimental results show that the given algorithm can determine the more reasonable clustering number of image segmentation and clustering center,meanwhile this method can realize image segmentation automatically in a fast and correct way.