Automated Characterization of Breast Cancer Using Steerable Filters
Hamido Fujita, U. Raghavendra, Anjan Gudigar, Vadakkepat Vinoy Vishnu, U. Rajendra Acharya · Frontiers in artificial intelligence and applications · 2017
Breast cancer is one of the highly researched topic in medical image analysis. Digital mammogram analysis is one of the techniques which helps in determining severity of breast cancer within the context of medical image analysis. In this work, a novel technique using steerable co-occurrence features and the independent component analysis (ICA) is proposed. Our method is evaluated using 1000 mammogram images and can efficiently classify normal, benign and malignant classes with a promising performance of 88.60% accuracy, using only ten features. The proposed method is completely automatic and it does not require any segmentation technique in the breast region.