Chaos-Based Biometrics Template Protection and Secure Authentication

With the increasing development of global economy and information technology, more and more fields require reliable identity authentication. And with information age characterized by digitalization and recessiveness of identity, a key problem to be solved is how to identify a person’s identity accurately and ensure information security. In this regard, a variety of inherent human biometrics were gradually understood and studied, thus the development of biometric identification technology is considerable. The gradual yet profound application of biometric identification system today has improved security and creates much convenience to identity authentication. However, there are still some inherent problems that need to be solved. For instance, masquerade attack, difficulties to republish when the template is lost and a series of other potential threats. The existences of these threats have created a bottleneck, constraining further development of the biometric identification technology. In this chapter, we will firstly give a review mainly on the theories and techniques of biometrics template protection, and then present a novel chaos-based biometrics template protection with secure authentication scheme. The proposed scheme is lightened by fuzzy extractor, yet includes two-layer error-correcting (one is BCH error-correcting code, the other is chaotic spread spectrum encryption) to achieve a good authentication performance of GAR=99.5% and FAR=0%. In addition, the functional features of proposed authentication scheme are: (1) do not need user to remember secret information such as password, or store them into physical media such as token or smart card; (2) no biometric template and any other secret information stored in server end; (3) the user’s biometric template is cancellable; (4) user’s registering information can be updated freely and easily. (5) with the help of user’s inaccurate biometric template, secret information (user maybe knows or unknowns) can be accurately recovered. These interesting features push forward the proposed scheme having potential application in biometric-based authentication/identification systems.