Orthogonal Discriminant Sparse Maximum Margin Analysis
Yue Lin, Jinlin Xu, Yu Zhang, Xingzhu Liang · DEStech Transactions on Engineering and Technology Research · 2017
Feature extraction is a crucial step for face recognition. In this paper, a new feature extraction method called orthogonal discriminant sparse maximum margin analysis (ODSMMA) is proposed for face recognition. ODSMMA defines two parameterless discriminant weighted matrices through the sparse representation. ODSMMA can efficiently preserve the margin between the same class and maximizes the margin between different classes simultaneously without setting any parameters. Moreover, by taking the advantage of the maximum margin criterion, ODSMMA is able to extract the orthogonal discriminant vectors in the feature space and does not suffer the small sample size problem. Experimental results on ORL database indicate the effectiveness of the proposed method.