The Application Research Of Mean Shift Algorithm In Mammograms Segmentation

Breast cancer is one of the diseases with high morbidity and mortality for women all over the world.Evidence has shown that the key to improving disease efficacy and reducing mortality is to diagnose and treat breast cancer patients as early as possible.Mammography which is convenient and low-cost,can identify the location and extent of lesions,so it is regared as the preferred method for breast cancer examination.However,mammography has some problems,such as low resolution,overlapping of breast structure and lesion location,so it is difficult to segment mammograms accurately.Therefore domestic and foreign scholars have conducted research on breast image segmentation deeply and proposed many segmentation methods.Its segmentation accuracy still needs to be improved due to the inherent defects of breast images.In order to segment the mass area accurately,Mean Shift clustering algorithm is applied to mammography images segmentation.Mean Shift algorithm is an unsupervised statistical classification algorithm,which does not need to specify the number of clusters in advance and does not depend on prior knowledge,so it is widely used in image segmentation,target tracking and other fields.Aiming at the problem of long iteration time and high time complexity ofMean Shift algorithm,this paper proposes to fuse Simple Linear Iterative Cluster(SLIC)algorithm with Mean Shift algorithm or combine Mean Shift algorithm with FLBP-Canny(Fuzzy Local Binary Pattern-Canny)algorithm,which fuses fuzzy local binary pattern and Canny edge,to segment breast images.(1)Aiming at the problem of the Mean Shift algorithm with many iterations and high time complexity,a new algorithm SLIC-MS base on SLIC is proposed.First,anisotropic diffusion and gamma correction are used to eliminate noise and enhance contrast because SLIC algorithm is very sensitive to noise.Then the image is pre-segmented into equalized super-pixel blocks by SLIC algorithm,the mean value of super-pixel blocks is calculated as the sample points of Mean Shift clustering algorithm,Mean Shift vector is calculated iteratively step by step.Finally,the threshold is obtained on the basis of the clustering density,and the clustering image is segmented by the threshold algorithm.Experiments show that the time complexity of SLIC-MS algorithm is reduced and its practicability can be enhanced to some extent.(2)In order to improve the accuracy of mammography images segmentation,a FLBP-Canny algorithm is proposed in this paper.Firstly,Mean Shift algorithm is used to filter the image,and then FLBP-Canny algorithm is used to extract the edge of breast mass to images segmentation.Experiments show that the segmentation accuracy of the algorithm reaches 90.26%,which is more conducive to mammography images segmentation.