Unimodal palmprint recognition system based on local features

Abstract

The most emerging biometric technology used to ensure a reliable recognition rate is Multibiometrics. The unimodal recognition systems may lead to low recognition rate in real applications. To overcome this problem, we propose an approach for palmprint recognition based on local features. First, a palmprint image is divided into several sub-images, then the feature vectors are extracted from each sub-block by uniform local binary pattern. The feature vectors of all the sub-images are combined together to form the feature vector. Finally the pattern classification can be assured by using classifiers based on Euclidian distance and City-block. The effectiveness of proposed method has been verified on PolyU palmprint database. The experimental results show that the recognition rates are significantly improved compared with others methods existing in literature. The recognition rate of the proposed method is the highest among the other algorithms. The optimal recognition rate obtained is 99,4%. The experimental results have shown that unimodal system is effective.