Robust Skin Color Driven Privacy Preserving Face Recognition via Function Secret Sharing

TL;DR

This paper introduces a privacy-preserving face recognition method that uses skin color features and a Function Secret Sharing protocol to enhance robustness against attacks and prevent information leakage.

Contribution

It presents a novel skin color feature extractor combined with an FSS-based face embedding comparison protocol, improving security and efficiency over existing methods.

Findings

Robust against black-box attacks and GAN-based image restoration.

More efficient than traditional Secret Sharing protocols.

Effectively prevents leakage of precomputed embeddings.

Abstract

In this work, we leverage the pure skin color patch from the face image as the additional information to train an auxiliary skin color feature extractor and face recognition model in parallel to improve performance of state-of-the-art (SOTA) privacy-preserving face recognition (PPFR) systems. Our solution is robust against black-box attacking and well-established generative adversarial network (GAN) based image restoration. We analyze the potential risk in previous work, where the proposed cosine similarity computation might directly leak the protected precomputed embedding stored on the server side. We propose a Function Secret Sharing (FSS) based face embedding comparison protocol without any intermediate result leakage. In addition, we show in experiments that the proposed protocol is more efficient compared to the Secret Sharing (SS) based protocol.