A High Performance and Secure Palmprint Template Protection Scheme

Abstract

Security and protection of biometric template has been the bottleneck of its applications due to permanent appearance of biometric features. To tackle this problem, many algorithms, that do not store the template in its original form, have been reported in recent years. In this paper, we propose a high performance and secure cancelable palmprint authentication algorithm for palmprint template protection based on texture features. Firstly, the anisotropic filter is employed to obtain the original palmprint template. Then, the feedforward–feedback nonlinear dynamic filter (FFNDF) is used to generate Cancelable Palmprint templates. At last, the independent matching scores are fused at the score level. The cancelable system can provide large re-issuance ability and can be implemented very fast for real applications. The experimental results on PolyU palmprint database show that both the chaotic key and biometrics play equally significant roles in cancelable and it can be easily generalized to other palmprint texture codes.