A Comparative Studyon Kernel PCA and PCA Methods for Face Recognition

Mohammad Mohsen Ahmadinejad, Elizabeth Sherly

Abstract: A online face recognition system is a dynamic topic in the fields of biometrics, which Many achievements have been obtained in face recognition. The human face has a principal role which, is consists of complicated combination of features that allows us to communicate, express our feelings and emotions. Principal Components Analysis (PCA) and kernel principal components, Analysis (KPCA) are achievements that, have been obtained in face feature extraction and recognition. In this papar have compared a PCA algorithm with KPCA algorithm, which a LFW data set is used for Compare, Recognition accuracy, Variation in Facial Expression, Illumination changes, and Computation time of each method. To find Recall of each algorithm are used a LFW database which shows a 2D-PCA have better performance.

Keywords: KPCA, PCA, algorithm, comparative, error rates