Research On Several Issues About Image Segmentation Based On Pulse Coupled Neural Network

Pulse Coupled Neural Network(PCNN) is a new neural network model which is known as the third generation of artificial neural networks. PCNN is based on the study of the mammalian visual cortex pulsation phenomena, so it has a biology background. PCNN is always a hot research field of digital image segmentation during the twenty years since it was used to image processing, because of its unique advantages. And PCNN also made a good research results in other intelligent information processing fields. In this paper, some problems exist when the pulse coupled neural network applications in the field of image segmentation is discussed and studied.Simplified pulse coupled neural network model achieved some good image segmentation results in the early years, but with the development of other segmentation techniques, its drawback gradually appeared, such as the edges and details of the deal is more rough and not sensitive to light and dark areas and so on. After in-depth study of the basic ideas and algorithm process of PCNN, and in order to improve the segmentation efficiency, we consider to combining the PCNN with the lateral inhibition mechanism which also has a biology background, and propose the Pulse Coupled Neural Network based on Lateral Inhibition(LI-PCNN) for image segmentation. Lateral inhibition mechanism can outstanding border fragment and clustering effect details. This method utilizes lateral inhibition factor regulating to the issuance of neurons adjacent pulses, imitating animal visual interaction between nerve cells, enhanced edge detail and shading treatment to achieve optimal segmentation results.The introduction of lateral inhibition mechanism had solved some of the problems in image segmentation, but it was found that the model parameters to have a significant impact of different segmentation results. The traditional method always has a great fortuitous for determine the parameters rely on experience and testing. Therefore, this paper is studied the PCNN parameters determine methods and propose Pulse Coupled Neural Network based on Fruit Fly Optimization Algorithm(FOA-PCNN), for adaptive parameter determination. The fruit fly optimization algorithm has excellent overall ability to find the optimal solution.This method avoids the blindness by using the FOA iterating find the optimal parameters of PCNN. The experiments in the MATLAB platform proved its effectiveness.Therefore, the main contents of this paper can be summarized as that: the research of the details by using the traditional PCNN model to image segmentation, and the study on the PCNN model parameters determination by using the swarm intelligence algorithm.