Two-factor cncellable biometrics use tokenized random number as an external factor, however, tokenized factor incurs severe security and privacy threats. In this paper, we propose a one-factor cancellable biometrics scheme, which requires sole biometric as input. First, it exploits an expanded feature vector, generating seeds of randomized binary auxiliary data by sliding a pre-defined window and Hash function, which enhances performance and security. Then, the cancellable template hence can be generated by replacing different binary auxiliary data. Finally, the auxiliary data is decoded by using sole query biometric. The experiments have been conducted on FVC2002 and FVC2004 databases, and results show that the scheme does not only fulfill four design criteria of cancellable biometrics but also resist to the threat model that enclosed three security attacks.
- Template protection