Robust and Secure Biometrics: Some Application Examples

In the past years there has been much theoretical interest in secure extraction of robust strings from noisy measurements. In particular this technique allows extracting robust cryptographic keys from noisy biometric data. This paper starts with an overview of the ideas behind robust and secure string extraction in terms of information reconciliation, privacy amplification and helper data. The main part of the paper gives three application examples explaining how these techniques are used to enhance the convenience and security of devices (e.g. tokens) while preserving also the privacy of the user because neither biometric information nor secret cryptographic keys need to be stored on applications. We discuss a server access token, a 3-way check for a biometric ePassport and a password vault.