Breast cancer classification using tetrolet transform based energy features and K-nearest neighbor classifier

Abstract

The cancer that develops in the breast tissue is referred to as breast cancer. The cancer in the breast could be sometimes symptomatic and are identified by self-examination of the breast or by a physician, whereas in certain cases there could be no symptoms at all. However, signs like a lump in the breast, change of size of the breast, dimpling and fluid discharge from the nipple are some of the symptoms which are cause of grave concern. The early diagnosis of the disease is the key to combat the deadly disease paving the path for hope of life. Mammography is a very popular technique that is used for the early diagnosis of breast cancer. In this study, a technique for breast cancer classification in digitized mammogram is put-forth employing tetrolet transform based energy features and K-Nearest Neighbor (KNN) classifier. The breast mammogram images of benign and malignant category are decomposed into sub-band coefficients using tetrolet transform and the energy features are extracted. These extracted features are given as input to the KNN classifier. Results show better classification accuracy in the breast cancer images using tetrolet transform based energy features and KNN classifier.