Cancelable Biometrics for HMM-based Signature Recognition

Along with the wide diffusion of biometric-based authentication systems, the need to provide security and privacy to the employed biometric templates has become an issue of paramount importance in the design of user-friendly applications. Unlike password or tokens, if a biometrics is compromised, usually it cannot be revoked or reissued. In this paper we propose an on-line signature-based biometric authentication system, where non invertible transformations are applied to the acquired signature functions, making impossible to derive the original biometrics from the stored templates, while maintaining the same recognition performances of an unprotected system. Specifically, the possibility of generating cancelable templates from the same original data, thus providing a proper solution to privacy concerns and security issues, is deeply investigated.