Reusable Fuzzy Extractor Based on the LPN Assumption

Abstract

Abstract A fuzzy extractor derives uniformly random strings from noisy sources that are neither reliably reproducible nor uniformly random. The basic definition of fuzzy extractor was first formally introduced by Dodis et al. and has achieved various applications in cryptographic systems. However, it has been proved that a fuzzy extractor could become totally insecure when the same noisy random source is extracted multiple times. To solve this problem, the reusable fuzzy extractor is proposed. In this paper, we propose the first reusable fuzzy extractor based on the LPN assumption, which is efficient and resilient to linear fraction of errors. Furthermore, our construction serves as an alternative post-quantum reusable fuzzy extractor.