Robustness of Facial Recognition to GAN-based Face-morphing Attacks

TL;DR

This paper investigates the robustness of facial recognition systems against GAN-based face-morphing attacks, revealing that improved FR algorithms can reduce attack success if morphs are included in threshold settings.

Contribution

It introduces two new GAN-based face-morphing attack methods and evaluates their effectiveness against current facial recognition algorithms.

Findings

GAN-based attacks can bypass existing FR systems.

Including morphed images in thresholds improves detection.

Enhanced FR algorithms reduce attack success rates.

Abstract

Face-morphing attacks have been a cause for concern for a number of years. Striving to remain one step ahead of attackers, researchers have proposed many methods of both creating and detecting morphed images. These detection methods, however, have generally proven to be inadequate. In this work we identify two new, GAN-based methods that an attacker may already have in his arsenal. Each method is evaluated against state-of-the-art facial recognition (FR) algorithms and we demonstrate that improvements to the fidelity of FR algorithms do lead to a reduction in the success rate of attacks provided morphed images are considered when setting operational acceptance thresholds.