An alignment free fingerprint fuzzy extractor using near-equivalent Dual Layer Structure Check (NeDLSC) algorithm

Biometric techniques provide a feasible solution for protecting personal information at a high security level. However, biometric features themselves demand protection. As a promising method, fuzzy extractor seamlessly binds a cryptographic key to biometric features, in order to prevent both keys and features being exposed to attackers. Some related research has been done in this area however three major issues are still unsolved, which are : 1) Feature alignment 2) Verification accuracy. 3) Information leakage. In this paper, an alignment-free fingerprint fuzzy extraction scheme is proposed. To eliminate the alignment process, we use the minutia local structure features which has been proved to be stable, discriminative, rotation and shift free. The Dual Layer Structure Check (DLSC) verification scheme is base on the minutia local structure and it achieves a high verification performance. In this paper, we proposed a near-equivalent version of DLSC (NeDLSC) that can be directly employed by the existing bio-cryptographic constructions. Then a new fuzzy extractor scheme which is on the basis of NeDLSC is demonstrated. The preliminary experiments on publicly-available database FVC2002 shows a high verification accuracy.