Enhanced template protection with passwords for fingerprint recognition

Template protection is an important supplementary to biometric systems for enhancing security and privacy protection. Its recognition and security performance is limited by inherent properties of the biometric modalities and the biometric systems used. Combining additional secret information such as PIN or password will be a promising way to improve the performance. The fuzzy vault is a widely-used cryptographic scheme to protect fingerprint minutiae. In this paper, we propose a novel method, which generates artificial minutiae from a PIN or password. A fused feature set including genuine and artificial minutiae is used to generate a protected template. The insertion of artificial minutiae increases the secret length as well as the robustness in the fuzzy vault. Our experimental results on the NIST SD 14 database show that both false match rate and false non-match rate are reduced in comparison with the original fuzzy vault method. Mean-while, the crucial security factors such as the degree of the polynomial and the length of secret are enlarged. Additionally, the original method is vulnerable to linkage attacks. The proposed method improves the resistance against this attack.