Backdoor Attacks on Deep Learning Face Detection

TL;DR

This paper investigates the vulnerability of face detection systems to object generation attacks, introduces a novel Landmark Shift Attack, and proposes mitigation strategies to enhance robustness against such backdoor threats.

Contribution

It presents the first Landmark Shift Attack on face detectors and offers mitigation methods, advancing understanding of backdoor vulnerabilities in face detection models.

Findings

Face Generation Attacks effectively deceive face detectors.

Landmark Shift Attack successfully backdoors coordinate regression.

Proposed mitigations improve robustness against these attacks.

Abstract

Face Recognition Systems that operate in unconstrained environments capture images under varying conditions,such as inconsistent lighting, or diverse face poses. These challenges require including a Face Detection module that regresses bounding boxes and landmark coordinates for proper Face Alignment. This paper shows the effectiveness of Object Generation Attacks on Face Detection, dubbed Face Generation Attacks, and demonstrates for the first time a Landmark Shift Attack that backdoors the coordinate regression task performed by face detectors. We then offer mitigations against these vulnerabilities.