A Multi-server Authentication Scheme Based on Fuzzy Extractor

Biometrics have been widely used in various authentication occasions, especially in the Internet field. However, the security of biometrics has attracted a lot of attention. Once the biometric template used for authentication is leaked, the user’s private data will face a great threat. As a key generation technology, fuzzy extractor can not only solve the problem of biometric template leaking, but can also realize the generation and regeneration of the secret keys. In previous research, we proposed an improved fuzzy extractor scheme (PrrFE). In this paper, a practical fuzzy extractor is applied to generate user’s Bio-ID for authentication along with a multi-server authentication scheme based on PrrFE. In this scheme, the users can use their own biometrics to generate a Bio-Key and then achieve authentication through the generated Bio-ID on more than one server. In our scheme, the biometric data and Bio-Key do not need to be stored anywhere. We have performed security analyses on our scheme. Our two experiments with the database FVC2002 and the NIST 302d have shown that the genuine acceptance rate (GAR) of real users is 99.16% in the optimal case. And it has a acceptable network delay.