Detection and classification of the breast abnormalities in Digital Mammograms via Linear Support Vector Machine
Nedra Amara, Muhammad Ali Shoaib, Said Gattoufi · 2018
This paper presents an approach to detect tumors in mammogram images. Early detection of breast cancer is key to scheming extremely good treatment strategies. The objective of this work is to distinguish between two classes of patients: those with benign or malignant tumor in Digital Mammograms via Linear Support Vector Machine classifier. The proposed methodology has been implemented in three steps: 1) estimation of an efficient k value selection for k-means segmentation of breast tissues; 2) using the fast and robust features descriptor Bag of Features based on SURF interest point for the extraction of features from the segmented tumor region; 3) the linear support vector machine classifier will be trained for the predication of tumor type into benign or malignant. The performance analysis of our proposal work is compared with three other state-of-the-art classifiers, BPN, KNN, and Hybrid RGSA. The experiments show that we succeeded to improve the accuracy for Benign and Malignant Breast tumors to 99.0%.