Template protection for on-line signature-based recognition systems

Security and privacy issues are considered as two of the major concerns related to the use of biometric data for authentication purposes. In this paper we propose two different approaches for the protection of on-line signature biometric templates. In the first one, cryptographic techniques are employed to protect signature features, making impossible to derive the original biometrics from the stored templates. In the second one, data hiding techniques are used to design a security scalable authentication system, embedding some dynamic signature features into a static representation of the signature itself. Extensive experimental results are provided to show the effectiveness of the presented protection methods. © 2008 Springer Berlin Heidelberg.