IEICE Information and Communication Technology Forum


Session Number:A2



Securing Practical Packing-based Privacy-preserving Biometric Authentication

Takumi Matsunaga,  Kentaroh Toyoda,  Iwao Sasase,  


Publication Date:2016-10-01

Online ISSN:2188-5079


PDF download (111.9KB)

Privacy-preserving biometric authentication is getting attractive. Recently, a not only privacy-preserving but also fast scheme has been proposed which is preferable for the real system. However, we point out that it has a flaw that an attacker can be authenticated with high probability. In this paper, we propose a secure practical packing-based privacy-preserving biometric authentication scheme which reduces the attack success rate. Our authentication scheme consists of three phases. At the first phase, the challenge feature vector is verified by the previous Hamming distance-based verification. In order to handle this vulnerability, we propose two more verifications which we call partial comparison and shift-then-inner product schemes. By combining three verifications, our scheme is secure against several attacking strategies. We evaluate the attack success rate, the true positive rate and the false positive rate with an iris dataset. As a result, we show that our proposed scheme can significantly reduce lower the attack success rate than the false positive rate while suppressively decreasing the true positive rate.