Real-world adversarial attack on MTCNN face detection system

TL;DR

This paper demonstrates a practical real-world adversarial attack on the MTCNN face detection system by using printed face attributes, highlighting security vulnerabilities in deep learning-based face detectors.

Contribution

The authors introduce a reproducible and robust method to attack MTCNN in real-world scenarios using printed face attributes attached to masks or faces.

Findings

The attack successfully fools MTCNN in real-world conditions.

Printed face attributes can be used to bypass face detection.

The method is easily reproducible with common printing tools.

Abstract

Recent studies proved that deep learning approaches achieve remarkable results on face detection task. On the other hand, the advances gave rise to a new problem associated with the security of the deep convolutional neural network models unveiling potential risks of DCNNs based applications. Even minor input changes in the digital domain can result in the network being fooled. It was shown then that some deep learning-based face detectors are prone to adversarial attacks not only in a digital domain but also in the real world. In the paper, we investigate the security of the well-known cascade CNN face detection system - MTCNN and introduce an easily reproducible and a robust way to attack it. We propose different face attributes printed on an ordinary white and black printer and attached either to the medical face mask or to the face directly. Our approach is capable of breaking the MTCNN detector in a real-world scenario.