Fast L1-eigenfaces for robust face recogntion
Matthew Johnson, Andreas E. Savakis · 2014
Face recognition using eigenfaces is a popular technique based on principal component analysis (PCA). However, its performance suffers from the presence of outliers due to occlusions and noise often encountered in unconstrained settings. We address this problem by utilizing L1-eigenfaces for robust face recognition. We introduce an effective approach for L1-eigenfaces based on combining fast computation of L1-PCA with a greedy search technique. Experimental results demonstrate that L1-eigenfaces outperform traditional L2-eigenfaces for face recognition and reconstruction on the Yale face database corrupted with random occlusions.