Mammogram-based discriminant fusion analysis for breast cancer diagnosis

Abstract

Mammogram-based classification is an important and effective way for computer-aided diagnosis (CAD)-based breast cancer diagnosis. In this paper, we present a novel discriminant fusing analysis (DFA)-based mammogram classification CAD-based breast cancer diagnosis. The discriminative breast tissue features are exacted and fused by DFA, and DFA achieves the optimal fusion coefficients. The largest class discriminant in the fused feature space is achieved by DFA for classification. Beside the detailed theory derivation, many experimental evaluations are implemented on Mammography Image Analysis Society mammogram database for breast cancer diagnosis.