On Brightness Agnostic Adversarial Examples Against Face Recognition Systems

TL;DR

This paper presents a new method for creating adversarial face images that remain effective despite brightness changes, improving the robustness of attacks on face recognition systems in real-world scenarios.

Contribution

A novel adversarial example generation technique that maintains effectiveness under brightness variations using non-linear transformations and curriculum learning.

Findings

Outperforms conventional methods in digital and physical tests.

Demonstrates robustness of adversarial examples against brightness changes.

Enables practical risk assessment of face recognition systems.

Abstract

This paper introduces a novel adversarial example generation method against face recognition systems (FRSs). An adversarial example (AX) is an image with deliberately crafted noise to cause incorrect predictions by a target system. The AXs generated from our method remain robust under real-world brightness changes. Our method performs non-linear brightness transformations while leveraging the concept of curriculum learning during the attack generation procedure. We demonstrate that our method outperforms conventional techniques from comprehensive experimental investigations in the digital and physical world. Furthermore, this method enables practical risk assessment of FRSs against brightness agnostic AXs.