Over-the-Air Adversarial Flickering Attacks against Video Recognition Networks

被引:25
|
作者
Pony, Roi [1 ]
Naeh, Itay [2 ]
Mannor, Shie [1 ,3 ]
机构
[1] Technion Israel Inst Technol, Dept Elect Engn, Haifa, Israel
[2] Rafael Adv Def Syst Ltd, Haifa, Israel
[3] Nvidia Res, Shanghai, Peoples R China
关键词
D O I
10.1109/CVPR46437.2021.00058
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Deep neural networks for video classification, just like image classification networks, may be subjected to adversarial manipulation. The main difference between image classifiers and video classifiers is that the latter usually use temporal information contained within the video. In this work we present a manipulation scheme for fooling video classifiers by introducing a flickering temporal perturbation that in some cases may be unnoticeable by human observers and is implementable in the real world. After demonstrating the manipulation of action classification of single videos, we generalize the procedure to make universal adversarial perturbation, achieving high fooling ratio. In addition, we generalize the universal perturbation and produce a temporal-invariant perturbation, which can be applied to the video without synchronizing the perturbation to the input. The attack was implemented on several target models and the transferability of the attack was demonstrated. These properties allow us to bridge the gap between simulated environment and real-world application, as will be demonstrated in this paper for the first time for an over-the-air flickering attack.
引用
收藏
页码:515 / 524
页数:10
相关论文
共 50 条
  • [1] Adversarial Deep Learning for Over-the-Air Spectrum Poisoning Attacks
    Sagduyu, Yalin E.
    Shi, Yi
    Erpek, Tugba
    [J]. IEEE TRANSACTIONS ON MOBILE COMPUTING, 2021, 20 (02) : 306 - 319
  • [2] Over-the-Air Adversarial Attacks on Deep Learning Wi-Fi Fingerprinting
    Xiao, Fei
    Huang, Yong
    Zuo, Yingying
    Kuang, Wei
    Wang, Wei
    [J]. IEEE INTERNET OF THINGS JOURNAL, 2023, 10 (11) : 9823 - 9835
  • [3] Over-the-Air Adversarial Attacks on Deep Learning Based Modulation Classifier over Wireless Channels
    Kim, Brian
    Sagduyu, Yalin E.
    Davaslioglu, Kemal
    Erpek, Tugba
    Ulukus, Sennur
    [J]. 2020 54TH ANNUAL CONFERENCE ON INFORMATION SCIENCES AND SYSTEMS (CISS), 2020, : 330 - 335
  • [4] IMPERIO: Robust Over-the-Air Adversarial Examples for Automatic Speech Recognition Systems
    Schoenherr, Lea
    Eisenhofer, Thorsten
    Zeiler, Steffen
    Holz, Thorsten
    Kolossa, Dorothea
    [J]. 36TH ANNUAL COMPUTER SECURITY APPLICATIONS CONFERENCE (ACSAC 2020), 2020, : 843 - 855
  • [5] Physical Layer Communications System Design Over-the-Air Using Adversarial Networks
    O'Shea, Timothy J.
    Roy, Tamoghna
    West, Nathan
    Hilburn, Benjamin C.
    [J]. 2018 26TH EUROPEAN SIGNAL PROCESSING CONFERENCE (EUSIPCO), 2018, : 529 - 532
  • [6] Communications Aware Adversarial Residual Networks for Over the Air Evasion Attacks
    Flowers, Bryse
    Buehrer, R. Michael
    Headley, William C.
    [J]. MILCOM 2019 - 2019 IEEE MILITARY COMMUNICATIONS CONFERENCE (MILCOM), 2019,
  • [7] One-Bit Aggregation for Over-the-Air Federated Learning Against Byzantine Attacks
    Miao, Yifan
    Ni, Wanli
    Tian, Hui
    [J]. IEEE SIGNAL PROCESSING LETTERS, 2024, 31 : 1024 - 1028
  • [8] EFFECTIVE AND INCONSPICUOUS OVER-THE-AIR ADVERSARIAL EXAMPLES WITH ADAPTIVE FILTERING
    O'Reilly, Patrick
    Awasthi, Pranjal
    Vijayaraghavan, Aravindan
    Pardo, Bryan
    [J]. 2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 6607 - 6611
  • [9] Practical Adversarial Attacks Against Speaker Recognition Systems
    Li, Zhuohang
    Shi, Cong
    Xie, Yi
    Liu, Jian
    Yuan, Bo
    Chen, Yingying
    [J]. PROCEEDINGS OF THE 21ST INTERNATIONAL WORKSHOP ON MOBILE COMPUTING SYSTEMS AND APPLICATIONS (HOTMOBILE'20), 2020, : 9 - 14
  • [10] Imperceptible adversarial attacks against traffic scene recognition
    Yinghui Zhu
    Yuzhen Jiang
    [J]. Soft Computing, 2021, 25 : 13069 - 13077