As a consequence of the wide deployment of biometrics-based recognition systems, there are increasing concerns about the security of the sensitive information managed. Various techniques have been proposed in the literature for the biometric templates protection (BTP), having gained great popularity the crypto-biometric systems. In the present paper we propose the implementation of a Fuzzy Vault (FV) scheme based on fixed-length templates with application to dynamic signature verification (DSV), where only 15 global features of the signature are considered to form the templates. The performance of the proposed system is evaluated using three databases: a proprietary collection of signatures, and the publicly available databases MCYT and BioSecure. The experimental results show very similar verification performance compared to an equivalent unprotected system.
Biometrics-based recognition systems have achieved great popularity in recent years. However, the widespread deployment of these systems has been affected by issues such as privacy and security. The theft of stored biometric information is considered one of the worst attacks on a biometric system. Several approaches have been developed to protect biometric templates, having gained great popularity the crypto-biometric constructions. In this article, a summary of the main biometric template protection approaches proposed is presented. In particular, the crypto-biometric Fuzzy Vault (FV) scheme is studied and some of its variants are analysed. Fuzzy Vault applications based on different modalities are reported.