On application of generalized Jacobi method in face recognition by linear discriminant analysis and principal component analysis
Vladimir V. Mokeyev · 2016
The paper considers efficiency of linear discriminant and principal component analysis regarding to “small sample size problems” for face recognition. Small sample size problem arises from a small number of available training samples compared to the dimensionality of the sample space. It results in singularity of the within-class scatter matrix. Different methods have been proposed to solve this problem in literature devoted to the face recognition issue. This study proposes calculation of the discriminant component by using the generalized Jacobi method, which prevents loss of discriminant information. High performance of the proposed approach is demonstrated in experimental studies based on the ORL database.