Cryptology ePrint Archive: Report 2018/1214

Instant Privacy-Preserving Biometric Authentication for Hamming Distance

Joohee Lee and Dongwoo Kim and Duhyeong Kim and Yongsoo Song and Junbum Shin and Jung Hee Cheon

Abstract: In recent years, there has been enormous research attention in privacy-preserving biometric authentication, which enables a user to verify him or herself to a server without disclosing raw biometric information. Since biometrics is irrevocable when exposed, it is very important to protect its privacy. In IEEE TIFS 2018, Zhou and Ren proposed a privacy-preserving user-centric biometric authentication scheme named PassBio, where the end-users encrypt their own templates, and the authentication server never sees the raw templates during the authentication phase. In their approach, it takes about 1 second to encrypt and compare 2000-bit templates based on Hamming distance on a laptop. However, this result is still far from practice because the size of templates used in commercialized products is much larger: according to NIST IREX IX report of 2018 which analyzed 46 iris recognition algorithms, size of their templates varies from 4,632-bit (579-byte) to 145,832-bit (18,229-byte).

In this paper, we propose a new privacy-preserving user-centric biometric authentication (HDM-PPBA) based on Hamming distance, which shows a big improvement in efficiency to the previous works. It is based on our new single-key function-hiding inner product encryption, which encrypts and computes the Hamming distance of 145,832-bit binary in about 0.3 seconds on Intel Core i5 2.9GHz CPU. We show that it satisfies simulation-based security under the hardness assumption of Learning with Errors (LWE) problem. The storage requirements, bandwidth and time complexity of HDM-PPBA depend linearly on the bit-length of biometrics, and it is applicable to any large templates used in NIST IREX IX report with high efficiency.

Category / Keywords: applications / privacy-preserving biometric authentication and inner product encryption and learning with errors

Date: received 18 Dec 2018, last revised 23 Dec 2018

Contact author: skfro6360 at snu ac kr

Available format(s): PDF | BibTeX Citation

Note: I fixed a typo in the author's name. (Jung Hee Cheon1 => Jung Hee Cheon)

Version: 20181224:025257 (All versions of this report)

Short URL:

[ Cryptology ePrint archive ]