Classification of Breast Abnormalities Using Deep Learning

—Early detection of breast abnormalities through mammography screening and proper treatment reduces mortality and increases women’s life expectancy. Currently, methods and algorithms for computer diagnostic systems based on deep neural networks are being actively developed. Such systems combine selection, feature calculation, and classification, thereby directly creating a decision-making function. In this paper, a method for classifying breast pathologies according to the Breast Imaging Reporting and Data System (BI-RADS) based on deep learning is proposed. Experimental results are presented using two open databases of digital mammography and evaluated using various performance criteria.