Palmprint identification using weighted PCA feature

As a feature extraction method, PCA has been wildly used in biometrics. Recently research shows that removing the first 3 eigenvectors can enhance the system performance for face recognition. In this paper, we investigate the influence by removing the first i eigenvectors of eigenspace firstly, then weighted PCA method is proposed, which has stronger ability than PCA under the same term. Meanwhile, it takes the best performance with fewer components without removing any bigger eigenvectors. Palmprint identification based on our database validates the algorithm.