A robust minutia-based approach for securing fingerprint templates

Abstract

Since the emergence of biometrics and its use in terms of authentication systems, the risks of fraud and identity theft keep increasing steadily. Hence, biometric template protection has become a real challenge for the research community. So several schemes have been proposed to address these issues and minimize the risk of attacks, but it turns out that just few of them have managed to satisfy both performance and security. In this context, we propose in our turn a fingerprint template protection, which consists of exploiting the minutiae structure that shapes the fingerprint, namely their positions and orientations, to then generate cancellable templates using a one-way transformation function. We proved that the proposal perfectly meets the properties of revocability, diversity, as well as accuracy performance. Promising results were obtained during an evaluation on FVC2002 DB1 database and which justifies the feasibility of our technique.