BadSAD: Clean-Label Backdoor Attacks against Deep Semi-Supervised Anomaly Detection
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912160214417408 |
|---|---|
| author | Cheng, He Xu, Depeng Yuan, Shuhan |
| author_facet | Cheng, He Xu, Depeng Yuan, Shuhan |
| contents | Image anomaly detection (IAD) is essential in applications such as industrial inspection, medical imaging, and security. Despite the progress achieved with deep learning models like Deep Semi-Supervised Anomaly Detection (DeepSAD), these models remain susceptible to backdoor attacks, presenting significant security challenges. In this paper, we introduce BadSAD, a novel backdoor attack framework specifically designed to target DeepSAD models. Our approach involves two key phases: trigger injection, where subtle triggers are embedded into normal images, and latent space manipulation, which positions and clusters the poisoned images near normal images to make the triggers appear benign. Extensive experiments on benchmark datasets validate the effectiveness of our attack strategy, highlighting the severe risks that backdoor attacks pose to deep learning-based anomaly detection systems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2412_13324 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | BadSAD: Clean-Label Backdoor Attacks against Deep Semi-Supervised Anomaly Detection Cheng, He Xu, Depeng Yuan, Shuhan Computer Vision and Pattern Recognition Artificial Intelligence Cryptography and Security I.2.6.e; I.5.4 Image anomaly detection (IAD) is essential in applications such as industrial inspection, medical imaging, and security. Despite the progress achieved with deep learning models like Deep Semi-Supervised Anomaly Detection (DeepSAD), these models remain susceptible to backdoor attacks, presenting significant security challenges. In this paper, we introduce BadSAD, a novel backdoor attack framework specifically designed to target DeepSAD models. Our approach involves two key phases: trigger injection, where subtle triggers are embedded into normal images, and latent space manipulation, which positions and clusters the poisoned images near normal images to make the triggers appear benign. Extensive experiments on benchmark datasets validate the effectiveness of our attack strategy, highlighting the severe risks that backdoor attacks pose to deep learning-based anomaly detection systems. |
| title | BadSAD: Clean-Label Backdoor Attacks against Deep Semi-Supervised Anomaly Detection |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Cryptography and Security I.2.6.e; I.5.4 |
| url | https://arxiv.org/abs/2412.13324 |