A New Method for Breast Cancer Identification Using Multi-modal Features in Quaternionic Form

Mammography is still the most effective procedure for early diagnosis of the breast cancer. Computer-aided Diagnosis (CAD) systems can be very helpful in this direction for radiologists to recognize abnormal and normal regions of interest in digital mammograms faster than traditional screening program. In this work, we propose a new method for breast cancer identification of all types of lesions in digital mammograms using multimodal features in a quaternionic representation. The proposed method consists of two steps: First, a novel feature extraction module utilizes two dimensional discrete transforms based on ART, Shapelets, Zernike moments and Gabor filters to decompose Regions of Suspicion (ROS) into a set of localized basis functions with different shapes. The extracted features are then fused and presented in quaternionic representation to the classification module in the second step. For the classification task, we propose a new type of classifier (Q-classifier) that successfully, accurately, with low computational cost and higher speed of diagnosis, recognizes normal and abnormal ROS from mammograms. The proposed method is evaluated on the Mini-MIAS database. The methods' performance is evaluated using Receiver Operating Characteristics (ROC) curve. The achieved result AUC = 0.934 shows that the proposed method can be quite effective and can be used as a tool for efficiently diagnosing breast cancer compared to similar techniques presented in the literature that use SVM classifiers and unimodal features.