TY - JOUR
T1 - Kernel machine-based one-parameter regularized fisher discriminant method for face recognition
AU - Chen, Wen Sheng
AU - Yuen, Pong C.
AU - Huang, Jian
AU - Dai, Dao Qing
N1 - Funding Information:
Manuscript received June 14, 2004; revised October 25, 2004. This work was supported by the Science Faculty Research Grant of Hong Kong Baptist University, RGC Earmarked Research Grant HKBU-2119/03E and NSFC (60144001, 10101013, 60175031, 10231040). This paper was recommended by Associate Editor I. Bloch.
PY - 2005/8
Y1 - 2005/8
N2 - This paper addresses two problems in linear discriminant analysis (LDA) of face recognition. The first one is the problem of recognition of human faces under pose and illumination variations. It is well known that the distribution of face images with different pose, illumination, and face expression is complex and nonlinear. The traditional linear methods, such as LDA, will not give a satisfactory performance. The second problem is the small sample size (S3) problem. This problem occurs when the number of training samples is smaller than the dimensionality of feature vector. In turn, the within-class scatter matrix will become singular. To overcome these limitations, this paper proposes a new kernel machine-based one-parameter regularized Fisher discriminant (K1PRFD) technique. K1PRFD is developed based on our previously developed one-parameter regularized discriminant analysis method and the well-known kernel approach. Therefore, K1PRFD consists of two parameters, namely the regularization parameter and kernel parameter. This paper further proposes a new method to determine the optimal kernel parameter in RBF kernel and regularized parameter in within-class scatter matrix simultaneously based on the conjugate gradient method. Three databases, namely FERET, Yale Group B, and CMU PIE, are selected for evaluation. The results are encouraging. Comparing with the existing LDA-based methods, the proposed method gives superior results.
AB - This paper addresses two problems in linear discriminant analysis (LDA) of face recognition. The first one is the problem of recognition of human faces under pose and illumination variations. It is well known that the distribution of face images with different pose, illumination, and face expression is complex and nonlinear. The traditional linear methods, such as LDA, will not give a satisfactory performance. The second problem is the small sample size (S3) problem. This problem occurs when the number of training samples is smaller than the dimensionality of feature vector. In turn, the within-class scatter matrix will become singular. To overcome these limitations, this paper proposes a new kernel machine-based one-parameter regularized Fisher discriminant (K1PRFD) technique. K1PRFD is developed based on our previously developed one-parameter regularized discriminant analysis method and the well-known kernel approach. Therefore, K1PRFD consists of two parameters, namely the regularization parameter and kernel parameter. This paper further proposes a new method to determine the optimal kernel parameter in RBF kernel and regularized parameter in within-class scatter matrix simultaneously based on the conjugate gradient method. Three databases, namely FERET, Yale Group B, and CMU PIE, are selected for evaluation. The results are encouraging. Comparing with the existing LDA-based methods, the proposed method gives superior results.
KW - Face recognition
KW - Pose and illumination variations
KW - RBF kernel function
KW - Regularized discriminant analysis
KW - Small sample-size problem
UR - http://www.scopus.com/inward/record.url?scp=24644515417&partnerID=8YFLogxK
U2 - 10.1109/TSMCB.2005.844596
DO - 10.1109/TSMCB.2005.844596
M3 - Journal article
C2 - 16128451
AN - SCOPUS:24644515417
SN - 1083-4419
VL - 35
SP - 659
EP - 669
JO - IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
JF - IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
IS - 4
ER -