A Survey on Fuzzy Private Set Intersection Protocols

Abstract

Similarity matching has been a fundamental operation in various real-world applications, e.g., illegal content detection and biometric authentication. In particular, regarding the increasing needs for maintaining data privacy during similarity matching, the fuzzy private set intersection (FPSI) protocols, which enable participants to privately learn which one party's set item is close enough to one of the counterpart's items, have attracted significant attention recently. In this paper, we survey existing FPSI protocols in various aspects, including distance metrics supported by protocols, protocol designs, and core cryptographic tools. Furthermore, we discuss (potential) technical challenges regarding deployments of them in real-world application scenarios.