A Face Based Fuzzy Vault Scheme for Secure Online Authentication

Biometric template protection is becoming a hot research recently. Fuzzy vault is a key binding biometric cryptosystem scheme. It binds a key with the biometric template and obtains the helper data. The main problem of fuzzy vault is that it can not provide the diversity and revocability. In this paper we propose a face based fuzzy vault scheme for online authentication. At the user side, the user's identity, the transformed template and key generated from user's password are always provided to the server. At the registration stage, fuzzy vault encoding is implemented using both key and transformed template. And then the fuzzy vault is encrypted using digital signature. At the authentication stage, by the identity claimed by user, the corresponding fuzzy vault will be checked, If it has been changed, the authentication will be failed. Otherwise, the key is recovered from the transformed template using fuzzy vault decoding, in which the nearest distance is computed for point matching. If the recovered key is same as that provided by user, the authentication will be successful. Otherwise it will be failed. The contribution of this paper include: the transformed template instead of face template is used, which will enable the system to provide diversity and revocability. The nearest distance based matching algorithm tolerates the much intra-class variation. The digital signature can be used to check if the fuzzy vault has been changed. The experimental results show that the proposed scheme achieves promising results.