Two-dimensional local graph embedding discriminant analysis (2DLGEDA) with its application to face and palm biometrics

This paper proposes a novel method, called two-dimensional local graph embedding discriminant analysis (2DLGEDA), for image feature extraction, which can directly extract the optimal projective vectors from two-dimensional image matrices rather than image vectors based on the scatter difference criterion. In graph embedding, the intrinsic graph characterizes the intraclass compactness and connects each data point with its neighboring within the same class, while the penalty graph connects the marginal points and characterizes the interclass separability. The proposed method effectively avoids the singularity problem frequently encountered in the traditional linear discriminant analysis algorithm (LDA) due to the small sample size (SSS) and overcomes the limitations of LDA due to data distribution assumptions and available projection directions. Experimental results on ORL, YALE, FERET face databases and PolyU palmprint database show the effectiveness of the proposed method.