An Improved Medical Image Segmentation Algorithm Based On Visual Perception Model

This paper is based on the review and sum up of the domestic and overseas literatures about the medical image segmentation. Analysis the classification and development thought of the segmented technology, We analysis the process of visual perception and attention mechanism based on the physiology and psychology of visual perception system. Meanwhile, according to the hierarchical characteristics of human visual perception, research two typical segmentation techniques of visual activity—the bottom—up segmentation and the up--bottom segmentation, According to its own characteristic and special application requirement of medical image, we start from the relationship between image segmentation and computer vision, and then propose a new composite method of medical image segmentation.Firstly, this method describes the underlying characteristic of image understanding, and then structures the layer characteristic, introduces local image texture features and describes the image area with visual perception consistency. Finally, it segments image by using Fuzzy-ART neural network as the classifier of pixels. The experimental results show that this medical image segmentation method can effectively solve the edge diluting problem of local information and overall distribution, achieving good visual effect of image segmentation.