Face Recognition Based on Uncorrelated Multilinear PCA Plus Classical LDA
Fan Zhang, Lin Qi, Enqing Chen · International Journal of Signal Processing Image Processing and Pattern Recognition · 2015
Subspace learning is an important direction in computer vision research.In this paper, a new method of face recognition based on uncorrelated multilinear principal component analysis (UMPCA) and linear discriminant analysis (LDA) is proposed.First, instead of transforming matrices into vectors for principal component analysis (PCA), UMPCA seeks a tensor-to-vector projection that captures most of the variation in the original tensorial input while producing uncorrelated features through successive variance maximization.A subset of features is extracted and the classical LDA is then applied to find the best subspaces.Finally, the comprehensive experiments are provided on AT&T databases and the experiment results show its superiority through the comparison with other PCA plus LDA based algorithms.