Support Vector Discriminant Analysis on Local Binary Patterns for Facial Expression Recognition
Zilu Ying, Lin-bo Cai · 2009
In this paper, a new approach to facial expression recognition is constructed by combining the support vector discriminant analysis (SVDA) and local binary pattern (LBP) operator. LBP is an effective low-cost image descriptor to extract facial texture representing expression features. The basic idea of SVDA is to find the projection axes according to the margin maximization criterion. SVDA is an excellent data dimension reduction method which benefits from the intrinsic merits of SVM such as generalization abilities and kernel tricks for nonlinear classification. The proposed algorithm is experimented on the Japanese female facial expression (JAFFE) database. Extensive experiments are carried out to compare with other common methods such as PCA and LDA. The experiment results show that the combination of LBP+SVDA provides better performance than that of those traditional algorithms and prove the effectiveness of the proposed algorithm.