Improved Facial Mask-Based Adversarial Attack for Deep Face Recognition Models

DOI

10.2991/978-94-6463-540-9_73How to use a DOI?

Keywords

Adversarial Attack; Face Recognition; Deep Learning

Abstract

This paper explores the enhancement of security and robustness in the field of facial recognition by investigating adversarial example attacks. The author not only introduces an advanced adversarial example generation technique by utilizing key facial landmarks, but also investigates universal mask-based adversarial example generation strategy. These research efforts increase the precision and efficiency of attacks and extend the scope, affecting a broader range of users. Through extensive experimental setups with the Residual Network (ResNet)-50 model and the Chinese Academy of Sciences (CASIA) Face Image Database Version 5.0 (CASIA-FaceV5), this paper assesses the effectiveness of the proposed methods under different attack scenarios and various evaluation criteria, such as L0, L1 norms, and the Structural Similarity Index. These results demonstrate that mask-based attacks and universal perturbations significantly reduce recognition accuracy while maintaining the concealment of the examples. This study emphasizes the security aspect of current facial recognition technology, which has profound implications for the safety of digital life.

Copyright

© 2024 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Haoran Wang
PY  - 2024
DA  - 2024/10/16
TI  - Improved Facial Mask-Based Adversarial Attack for Deep Face Recognition Models
BT  - Proceedings of the 2024 2nd International Conference on Image, Algorithms and Artificial Intelligence (ICIAAI 2024)
PB  - Atlantis Press
SP  - 714
EP  - 722
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-540-9_73
DO  - 10.2991/978-94-6463-540-9_73
ID  - Wang2024
ER  -