Diagonal principal component analysis with non-greedy ℓ1-norm maximization for face recognition

Diagonal principal component analysis (DiaPCA) is an important method for dimensionality reduction and feature extraction. It usually makes use of the ℓ2-norm criterion for optimization, and is thus sensitive to outliers. In this paper, we present a DiaPCA with non-greedy ℓ1-norm maximization (DiaPCA-L1 non-greedy), which is more robust to outliers. Experimental results on two benchmark datasets show the effectiveness and advantages of our proposed method.