How to use ART to generate dodging or impersonation attack adversarial examples on face recognition task？

[PoseidomWong]

How to use ART to generate dodging or impersonation attack adversarial examples on face recognition task？Anyone kowns?

[beat-buesser]

Hi @PoseidomWong Can you tell us more about the impersonation attack (maybe with a paper link)?

[PoseidomWong]

Like this work:
https://github.com/ShawnXYang/Face-Robustness-Benchmark
And in this paper also has related introduction:
https://www.sciencedirect.com/science/article/abs/pii/S1077314220301296
Thank you!

[beat-buesser]

Many of the attacks in this work are available in ART (e.g. FGSM, BIM, C&W). You could follow similar steps as in our example scripts get_started_*.py in https://github.com/Trusted-AI/adversarial-robustness-toolbox/tree/main/examples but for the face datasets and models.