Occluded Face Recognition Using Sparse Complex Matrix Factorization with Ridge Regularization

Abstract

Matrix factorization is a method for dimensionality reduction which plays an important role in pattern recognition and data analysis. This work exploits the usefulness of our proposed complex matrix factorization (CMF) with ridge regularization (SCMF-L2) in occluded face recognition. Experiments on occluded face recognition reveal that the SCMF-L2 method provides the best recognition result among all the nonnegative matrix factorization (NMF) and CMF methods. The proposed method also reaches the stopping condition and converge much faster than the other NMF and CMF methods.