Generating Master Faces for Use in Performing Wolf Attacks on Face Recognition Systems

TL;DR

This paper demonstrates that 'master faces' generated using StyleGAN can successfully perform wolf attacks on face recognition systems, highlighting vulnerabilities and suggesting ways to improve system robustness.

Contribution

It introduces the concept of master faces for face recognition attacks and shows how to generate them using state-of-the-art face synthesis techniques.

Findings

Master faces can match many enrolled templates, enabling wolf attacks.

Pre-trained models are sufficient for effective master face generation.

The work highlights vulnerabilities and potential improvements in face recognition security.

Abstract

Due to its convenience, biometric authentication, especial face authentication, has become increasingly mainstream and thus is now a prime target for attackers. Presentation attacks and face morphing are typical types of attack. Previous research has shown that finger-vein- and fingerprint-based authentication methods are susceptible to wolf attacks, in which a wolf sample matches many enrolled user templates. In this work, we demonstrated that wolf (generic) faces, which we call "master faces," can also compromise face recognition systems and that the master face concept can be generalized in some cases. Motivated by recent similar work in the fingerprint domain, we generated high-quality master faces by using the state-of-the-art face generator StyleGAN in a process called latent variable evolution. Experiments demonstrated that even attackers with limited resources using only pre-trained models available on the Internet can initiate master face attacks. The results, in addition to demonstrating performance from the attacker's point of view, can also be used to clarify and improve the performance of face recognition systems and harden face authentication systems.