Cryptology ePrint Archive: Report 2018/681

A Reusable Fuzzy Extractor with Practical Storage Size

Jung Hee Cheon and Jinhyuck Jeong and Dongwoo Kim and Jongchan Lee

Abstract: After the concept of a Fuzzy Extractor (FE) was rst introduced by Dodis et al. , it has been regarded as one of the candidate solutions for key management utilizing biometric data. With a noisy input such as biometrics, FE generates a public helper value and a random secret key which is reproducible given another input similar to the original input. However, "helper values" may cause some leakage of information when generated repeatedly by correlated inputs, thus reusability should be considered as an important property. Recently, Canetti et al. (Eurocrypt 2016) proposed a FE satisfying both reusability and robustness with inputs from low-entropy distributions. Their strategy, the so-called Sample-then-Lock method, is to sample many partial strings from a noisy input string and to lock one secret key with each partial string independently. In this paper, modifying this reusable FE, we propose a new FE with size-reduced helper data hiring a threshold scheme. Our new FE also satis es both reusability and robustness, and requires much less storage memory than the original. To show the advantages of this scheme, we analyze and compare our scheme with the original in concrete parameters of the biometric, IrisCode. As a result, on 1024-bit inputs, with false rejection rate 0.5 and error tolerance 0.25, while the original requires about 1TB for each helper value, our scheme requires only 300MB with an additional 1.35GB of common data which can be used for all helper values.

Category / Keywords: secret-key cryptography / threshold cryptography, key management, secret sharing, applications, fuzzy extractor, biometric authentication

Original Publication (with minor differences): Australasian Conference on Information Security and Privacy (ACISP) 2018
DOI:
10.1007/978-3-319-93638-3_3

Date: received 16 Jul 2018

Contact author: wlsyrlekd at snu ac kr

Available format(s): PDF | BibTeX Citation

Version: 20180716:140010 (All versions of this report)

Short URL: ia.cr/2018/681


[ Cryptology ePrint archive ]