Efficient privacy-preserving face verification scheme

Abstract

Face verification has been a promising artificial intelligence technique and is becoming increasingly popular in civilian and law enforcement applications that require reliable biometric authentication. However, the wide deployment of face verification systems has raised serious privacy concerns regarding the abuse of users’ face images. In this paper, we present an efficient privacy-preserving face verification scheme based on fully homomorphic encryption and garbled circuit. Our protocol is the first one that guarantees that the server cannot learn anything about users’ face images and the verification result while providing the face verification service, and the client learns nothing beyond the verification result. Experimental results show that our scheme takes less than one second to perform a privacy-preserving face verification.