Cancelable Voiceprint Templates Based on Knowledge Signatures

Abstract

Biometric-based authentication can provide strong safety guarantee of user identity, but creates other concerns pertaining to template security. How to prevent the templates from being abused, and how to protect the users' privacy well are important problems. This paper proposes an approach based on knowledge signatures to solve these problems, which provides a cancelable template. Only a non-invertible transformed version of the biometrics is stored in the server, and the original data can't be obtained, so the users' privacy can be protected well. Moreover, distinguished with other cancelable templates, it only needs an incompact association between the key and the biometrics feature, even if the key is disclosed, the original biometrics feature wouldn't be revealed. The experimental results show that the proposed approach is practical.