A Quasilinear-Time CVP Algorithm for Triangular Lattice Based Fuzzy Extractors and Fuzzy Signatures

Fuzzy Extractor (FE) and Fuzzy Signature (FS) are useful schemes for generating cryptographic keys from fuzzy data such as biometric features. To implement FE and FS for fuzzy data in a Euclidean space, such as facial feature vectors, several techniques have been proposed that use lattice-based error correction. In these techniques, it is necessary to find the closest vector in the lattice to a given fuzzy data at the time of key reproduction or signing. However, the closest vector problem (CVP) in a general lattice is known to be NP-hard. Therefore, it is necessary to use a special lattice in which an efficient CVP algorithm can be constructed. In this paper, we propose a faster algorithm for CVP in triangular lattices with (nlogn)-time compared to the conventional O(n²)-time algorithm.