EUVIMEDEUVIMED entdecken
Crossref

Pre-trained Trojan Attacks for Visual Recognition

Aishan Liu, Xianglong Liu, Xinwei Zhang, Yisong Xiao, Yuguang Zhou, Siyuan Liang, Jiakai Wang, Xiaochun Cao, Dacheng Tao

International Journal of Computer Vision · 2025-06 · 133 (6) · 3568-3585

DOI: 10.1007/s11263-024-02333-y ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Computer Vision, online erschienen am 22.01.2025. Verfasst von Aishan Liu und 8 weiteren Autor:innen. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 63 Literaturangaben. Gefördert von: National Key R&D Program of China, National Natural Science Foundation of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Pre-trained vision models (PVMs) have become a dominant component due to their exceptional performance when fine-tuned for downstream tasks. However, the presence of backdoors within PVMs poses significant threats. Unfortunately, existing studies primarily focus on backdooring PVMs for the classification task, neglecting potential inherited backdoors in downstream tasks such as detection and segmentation. In this paper, we propose the Pre-trained Trojan attack, which embeds backdoors into a PVM, enabling attacks across various downstream vision tasks. We highlight the challenges posed by cross-task activation and shortcut connections in successful backdoor attacks. To achieve effective trigger activation in diverse tasks, we stylize the backdoor trigger patterns with class-specific textures, enhancing the recognition of task-irrelevant low-level features associated with the target class in the trigger pattern. Moreover, we address the issue of shortcut connections by introducing a context-free learning pipeline for poison training. In this approach, triggers without contextual backgrounds are directly utilized as training data, diverging from the conventional use of clean images. Consequently, we establish a direct shortcut from the trigger to the target class, mitigating the shortcut connection issue. We conducted extensive experiments to thoroughly validate the effectiveness of our attacks on downstream detection and segmentation tasks. Additionally, we showcase the potential of our approach in more practical scenarios, including large vision models and 3D object detection in autonomous driving. This paper aims to raise awareness of the potential threats associated with applying PVMs in practical scenarios. Our codes are available at https://github.com/Veee9/Pre-trained-Trojan.

Förderung

  • National Key R&D Program of China
  • National Natural Science Foundation of China

Zitieren

APA

Liu, A., Liu, X., Zhang, X., Xiao, Y., Zhou, Y., Liang, S., Wang, J., Cao, X., & Tao, D. (2025). Pre-trained Trojan Attacks for Visual Recognition. International Journal of Computer Vision, 133(6), 3568-3585. https://doi.org/10.1007/s11263-024-02333-y

Vancouver

Liu A, Liu X, Zhang X, Xiao Y, Zhou Y, Liang S, et al. Pre-trained Trojan Attacks for Visual Recognition. International Journal of Computer Vision. 2025;133(6):3568-3585. doi:10.1007/s11263-024-02333-y

BibTeX
@article{Liu2025,
  author = {Liu, Aishan and Liu, Xianglong and Zhang, Xinwei and Xiao, Yisong and Zhou, Yuguang and Liang, Siyuan and Wang, Jiakai and Cao, Xiaochun and Tao, Dacheng},
  title = {Pre-trained Trojan Attacks for Visual Recognition},
  journal = {International Journal of Computer Vision},
  year = {2025},
  volume = {133},
  number = {6},
  pages = {3568--3585},
  doi = {10.1007/s11263-024-02333-y}
}
Kostenlos starten →