A Novel Discriminant Criterion with Application to Face Recognition

Linear discriminant analysis(LDA) is one of the most popular methods for feature extraction and dimensionality reduction, but it may encounter the so called small sample size(SSS) problem when applied to high dimensional data analysis such as face recognition. In addition, the between-class scatter matrix defined in terms of class centroids places no restrictions on individual samples, maximization of which will cause a large overlap of neighboring classes(NCO). In this paper, we propose a novel discriminant criterion including two different measures that can solve the SSS and NCO problem, in which the optimal discriminant vectors are the eigenvectors of the weighted sum of within-class scatter and between-class similarity matrix corresponding to the minimum eigenvalues. In terms of classification performance, the proposed method outperforms or compares favorably several state-of-the-art methods. The effectiveness of our method is verified in the experiments on some benchmark face database.