An Improved Face Recognition Fusion Algorithm Based on the Features extracted from Gabor, PCA and KPCA

Abstract

The efficiency of face recognition can be affected by many factors such as illumination, posture, occlusion and so on, which can be summarized as the combination of linear and nonlinear interference variables. To solve this problem, this paper proposes a face recognition algorithm based on the Gabor wavelet, principal component analysis (PCA) and kernel PCA (KPCA). Specifically, bilinear interpolation is introduced to preprocess the original face database. Then, Gabor wavelets are used to extract the detailed features of the faces. The extracted Gabor features are dimensional reduced by two methods, respectively, which are: principal component analysis (PCA) plus linear discriminant analysis (LDA) and KPCA plus LDA. The two dimensional reduced features are then fused together with certain weights. All the experiments are based on the YALE face database and the ORL face database. The proposed algorithm can effectively improve the efficiency of face recognition in the complex environment.