K. Vaidehi ¹, T. S. Subashini ¹

Affiliations
1 Department of Computer Science and Engineering, Annamalai University, Annamalai Nagar, 608002, IN

Abstract

Worldwide, breast cancer is one of the top two lethal diseases among women. Breast tissue density is the important risk indicator of breast cancer. Digital Mammography technique is used to detect the breast cancer at its benign stage. Computer Aided Diagnosis (CAD) tools aids the radiologist for an accurate diagnosis and interpretation. In this work, Statistical features are extracted from the Region of Interest (ROI) of the breast parenchymal region. K-NN with three different distance metrics namely Euclidean, Cosine, City-block and its combination is used for classification. The extracted features are fed into the classifier to classify the ROI into any of three breast tissue classes such as dense, fatty, glandular. The classification accuracy obtained for combined k-NN is 91.16%.

Keywords

Breast Density, K-NN, Mammography, Statistical Descriptors.

Full Text