A Novel Feature Extraction Method Based on 2-D ARMA Model for Classification of Breast Cancer Images

Abstract—This Using Computer aided diagnosis (CAD) systems is becoming popular for discriminating malignant from benign lesions in MRI breast images. Different CAD system rely on different methods of feature extraction and classification. Extracting a more efficient features can result in better discrimination rate employing less computational power. In this paper, we propose a CAD system for tumor classification (cancerous vs. benign) in MRI breast images based on a twodimensional (2-D) autoregressive moving average (ARMA) model of the breast image. Then, the estimated parameters are used as the basis for statistical inference and biophysical interpretation of the breast image. We evaluate the performance of the 2D ARMA vector features in real MRI images using a SVM classifier. Our results suggest that the proposed CAD system based on a 2-D ARMA model leads to parameters that can accurately classify the MRI breast image into two regions: benign tumor, and cancerous tumor. Moreover, the extracted features are rich enough to discriminate between the two groups using a linear classifier.