This directory contains benchmark face databases (AR, FERET, GT, ORL, and Yale) used to evaluate our proposed RSSPCA algorithm in comparison with established methods including PCA, PCA-L1, and RSPCA.