Research And Application On The Algorithm Of Medical Image Segmention

In recent years medical image segmentation technology is one of the important subjects in medical image processing and analysis research field, and has been a hot issue for researchers.The main purpose of medical image segmentation is to divide the image into different regions with special signification and to make the results approximate to the anatomic structure, which can provide reliable basis for diagnosis and treatment and pathology research. As the complexity of human anatomic structure,applying a single image segmentation method to medical image segmentation can not get ideal results. Therefore, effective methods must be found to resolve the problem.This paper reviews and summarizes medical image segmentation method at home and abroad, and explaines the purpose and significance of medical image segmentation.According to the characteristics of medical images and application demand, an integrated medical image segmentation method is proposed, which joints fuzzy c-means based on Gibbs random field with snake model. At the same time, this paper investigates the algorithm of the level set method which needn't to re-initialize in the program. The main contents are as follows:1. This paper discusses popular image segmentation methods at home and abroad, overviewes the basic concepts, basic theory and basic methods of medical image segmentation technology, and also classifies today's medical image segmentation algorithms.2. According to the characteristics of medical images, applying a single image segmentation method to medical image segmentation can not get ideal results. This paper proposes an integrated medical image segmentation method is proposed, which joints fuzzy c-means based on Gibbs random field with snake model.We apply this method to the epi-metaphyseal regions of interest to obtaine correct location on the edge of phalanges. As the distance between epiphysis and metaphysis is relatively close, it can prevent the contour from closing toward the edge of epiphyseal structure which has been divided in the first step, to set a perpendicular to the x-axis of region of interest in the segmentation process. Experiments have proved that meaningful improvement in the epiphysis of the bone at the bottom of the convex region has been obtained. At the same time the joint approach is more immune to noise.3. Genetic algorithm is introduced to the level set method of no need to re-initialize in image segmentation, which solves the problems of the level set method, such as huge search space of parameters and time-consuming.In order to assess the parameters of level set method, we use fuzzy C-means method to obtain the expectations border, and compare it with the result contours of the level set methods to get a set of optimal parameters.This method improves the efficiency and accuracy of algorithm, and can also be applied to other areas of image processing.