A face recognition algorithm based on global and local feature fusion
Yi YH, Qu DK, Xu F(徐方) · SIA OpenIR (Chinese Academy of Sciences) · 2010
Numerous studies in psychophysics and neurophysiological literatures have shown that both local and global features are important for representing and recognizing face. In this paper, we propose an noval face representation and recognition approach which combining global and local discriminative features efficiently. In our method, global features are extracted from whole face images by LDA and local features are extracted by Gabor wavelet transform and weighted according to importance of local region. We generate two classifiers by taking full advantage of bayesian rule, all these classifiers are combined to form a hierarchical ensemble. We evaluated the proposed method compared with other classical algorithms. Experiments on the IMDB databases show that our method achieves satisfactory performance not only under the conditions of varied facial expression and lighting configuration but also under the conditions where the pose and sample size are varied.