Fusing the complete linear discriminant features by fuzzy integral for face recognition

Abstract

A complete linear discriminant analysis (CLDA) algorithm is proposed in this paper, which can extract the discriminant features both in the null space and the range space. Based on CLDA, a three-phase framework is proposed for face recognition. A face image is firstly decomposed by wavelet transform and its global and local information is obtained. Secondly, CLDA is used to extract the complete discriminant features contained in the global and local information. Finally, These different kinds of information are fused by fuzzy integral for the purpose of classification. The experimental results demonstrate that the proposed method yields better classification performance in comparison to the results obtained by other methods, such as Eigenface, Fisherface, KPCA or KFD methods