Cancelable palm print templates using Gabor representations and random measure

Abstract

Palm print template security is an important issue in a real biometric system because compromised templates cannot be revoked and reissued. To tackle these problems, in this paper, by measuring the Gabor-based palm print image with a chaotic random matrix, we propose a palm print template protection scheme with provable security and acceptable recognition performance. Firstly, the Gabor is employed to convolute with preprocessed palm print image. Then, to generate the binary cancelable template, the obtained Gabor representation is compared with a chaotic random matrix. The random measure can improve the template's discriminability. Therefore, the proposed algorithm not only protects the template but also has better performance. The experimental results on Hong Kong PolyU palm print database show that the proposed approach can achieve zero equal error rate (EER) and has large cancel ability. © 2013 Springer-Verlag.