Application Of Fuzzy Clustering Algorithm In Image Segmentation

In the computer-aided diagnosis systems based on medical images,accurate segmentation of abnormal regions in medical images is a crucial step.Breast cancer is a high incidence disease of modern women.Breast mammography is the main method for breast cancer screening.It plays an important role in the early detection and early treatment of breast cancer and the reduction of breast cancer mortality.The molybdenum target image of mammography usually has low resolution,and the difference in gray level between the abnormal region and the glandular tissue is small,and there are some overlap and interweave.Therefore,it is difficult to accurately segment abnormal regions in mammography images.At present,there are many methods for segmenting breast masses,but the focus of various methods is different.It is still necessary to further optimize the segmentation algorithm to improve the accuracy of segmentation.Fuzzy clustering algorithm is an image segmentation method that introduces fuzzy theory.Among the fuzzy clustering algorithms,the most widely used one is the Fuzzy C-Means(FCM)algorithm based on the objective function.Since fuzzy C-means clustering algorithm is linearly inseparable,related scholars have proposed Kernel Fuzzy C-Means(KFCM)algorithm to achieve effective clustering of various data structures.In this thesis,the kernel fuzzy C-means clustering algorithm is optimized for the deficiency,and combined with other image segmentation methods for breast mammography image segmentation.Firstly,random initialization of the cluster center will lead to unstable clustering results.Concerning this issue,an improved kernel fuzzy C-means clustering algorithm based on improved bat algorithm is proposed.The improved bat algorithm adjusts the initialization method and flight mode of the traditional bat algorithm to find the optimal cluster center set,and then the clustering is done by the kernel fuzzy c-means clustering algorithm.Secondly,the two-dimensional histogram of the image is combined with the optimized kernel fuzzy C-means clustering algorithm for segmentation of mammography images.According to the two-dimensional histogram of gray-gradient of the preprocessed image,the clustering number of the image is determined,and then the optimized kernel fuzzy c-means clustering algorithm is used to segment the mass.To verify the performance of the algorithm,the optimized kernel fuzzy C-means clustering algorithm was tested on the common data set.It is proved that the algorithm proposed in this paper achieves higher time efficiency with higher clustering accuracy.In addition,when the algorithm combined with the two-dimensional histogram is used for breast molybdenum target imagesegmentation,over-segmentation of the traditional fuzzy clustering algorithm is avoided,and the region of interest can be segmented more accurately.