Medical Image Segmentation Based On Fuzzy Clustering Method

Image segmentation is a basic computer vision technology, the main task of image segmentation is to divide an image into different regions, and extract the interesting regions.Medical image segmentation is an important image segmentation application. Image segmentation is a classical difficult problem in medical image processing and analysis. It is the basis of clinical trials, lesion region extraction, measurement of specific organizations, as well as the basis realization of three-dimensional reconstruction. It has important significance in medical research and clinical applications.In this paper, we focus on fuzzy clustering methods, and we analyze the principle of fuzzy clustering algorithm in medical image segmentation. The main contents of this article include the following:(1) Introduce several kinds medical image segmentation methods, they are segmentation methods based on threshold, methods based on region characteristics, methods based on edge detection, and methods based on some kinds of specific theory. And we use some segmentation methods to segment medical images.(2) In this paper, we focus on fuzzy clustering methods, and we analyze the principle of fuzzy clustering algorithm. because of the shortcomings of the fuzzy c-means methods, we introduce a new method which uses a difference control function to modify the membership of the spatial information, and the new membership gives different measurements to the edge areas, smaller areas , noise with the inside areas. the results show that the proposed algorithm can provide more better results of noisy medical images than fuzzy c-means algorithm and spatial fuzzy c-means algorithm, and the proposed method has fewer iterations.(3) Introduce a medical image segmentation method which combines the fuzzy c-means clustering methods with the level set segmentation method. Some traditional level set methods, such as C-V model method do not using the gradient information, and they need re-initialization, so they are low efficiency methods. In this paper, we use a level set evolution without re-initialization method. The method needs not re-initialization, and it uses the gradient information. So it has good results for the clear edge objects. But it can not get ideal results when the edge is ambiguous and when the edge of the image contour lines may be prone to weak border and cross the border or stop where there are still some distance to actual edges. In this paper, we combine fuzzy c-means clustering methods with level set evolution without re-initialization method. Firstly, use the fuzzy c-means method to class the image into different classes, this step can makes larger gray-scaledifference among the different parts of the image, and then use the level set evolution withoutre-initialization method to segment images. The results show that the proposed method is a bettermedical image segmentation method.