Segmentation of Mammogram Using Tumor-Cut Algorithm
V. S. Gowri · 2013
Mammography is the most effective technique used by radiologist for the screening and diagnosis of breast cancer. The detection rate and accuracy of breast cancer in mammograms depend on the segmentation of images. Many segmentation algorithms like Watershed, Region-Growing, K-means Clustering, Edge detection etc are used for the detection of tumors and their advantages and disadvantages are discussed in this paper. Even then the detection rate is still not high. In the proposed method a Tumor-cut algorithm is used for the segmentation of mammogram to increase the detection rate. Noise and artifact removal is performed with the help of Gabor filter. Features are extracted from the segmented images and classified as benign and malignant using support vector machine. To demonstrate the proposed approach 20 digital mammogram images are taken from DDSM (Digital Database for Screening Mammography) database.