GPCA on Gabor Tensor for Face Recognition

Lian Zhu,Rong Huang,Xinfu Ye,Wankou Yang,Sun Changyin
DOI: https://doi.org/10.1007/978-3-642-42057-3_44
2013-01-01
Abstract:There is a growing interest in subspace learning techniques for face recognition such as Gabor face representation. Although the Gabor face representation has received great success in face recognition, the excessive dimension of the data space often brings the algorithms into the curse of dimensionality dilemma. This paper proposes a novel face recognition method based on discriminant analysis with Gabor tensor representation. We derive a 3rd-order Gabor tensor representation from a complete response set of 40 Gabor filters. Then Generalized Principal Component Analysis (GPCA) is applied to each Gabor feature matrix. After working out 40 times, the feature matrices in a lower dimensional subspace are finally integrated for classification. Experimental results on ORL database and AR database show promising results of the proposed method.
What problem does this paper attempt to address?