BadDet+: Robust Backdoor Attacks for Object Detection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912857719832576 |
|---|---|
| author | Dunnett, Kealan Arablouei, Reza Miller, Dimity Dedeoglu, Volkan Jurdak, Raja |
| author_facet | Dunnett, Kealan Arablouei, Reza Miller, Dimity Dedeoglu, Volkan Jurdak, Raja |
| contents | Backdoor attacks pose a severe threat to deep learning, yet their impact on object detection remains poorly understood compared to image classification. While attacks have been proposed, we identify critical weaknesses in existing detection-based methods, specifically their reliance on unrealistic assumptions and a lack of physical validation. To bridge this gap, we introduce BadDet+, a penalty-based framework that unifies Region Misclassification Attacks (RMA) and Object Disappearance Attacks (ODA). The core mechanism utilizes a log-barrier penalty to suppress true-class predictions for triggered inputs, resulting in (i) position and scale invariance, and (ii) enhanced physical robustness. On real-world benchmarks, BadDet+ achieves superior synthetic-to-physical transfer compared to existing RMA and ODA baselines while preserving clean performance. Theoretical analysis confirms the proposed penalty acts within a trigger-specific feature subspace, reliably inducing attacks without degrading standard inference. These results highlight significant vulnerabilities in object detection and the necessity for specialized defenses. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_21066 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | BadDet+: Robust Backdoor Attacks for Object Detection Dunnett, Kealan Arablouei, Reza Miller, Dimity Dedeoglu, Volkan Jurdak, Raja Computer Vision and Pattern Recognition Cryptography and Security Backdoor attacks pose a severe threat to deep learning, yet their impact on object detection remains poorly understood compared to image classification. While attacks have been proposed, we identify critical weaknesses in existing detection-based methods, specifically their reliance on unrealistic assumptions and a lack of physical validation. To bridge this gap, we introduce BadDet+, a penalty-based framework that unifies Region Misclassification Attacks (RMA) and Object Disappearance Attacks (ODA). The core mechanism utilizes a log-barrier penalty to suppress true-class predictions for triggered inputs, resulting in (i) position and scale invariance, and (ii) enhanced physical robustness. On real-world benchmarks, BadDet+ achieves superior synthetic-to-physical transfer compared to existing RMA and ODA baselines while preserving clean performance. Theoretical analysis confirms the proposed penalty acts within a trigger-specific feature subspace, reliably inducing attacks without degrading standard inference. These results highlight significant vulnerabilities in object detection and the necessity for specialized defenses. |
| title | BadDet+: Robust Backdoor Attacks for Object Detection |
| topic | Computer Vision and Pattern Recognition Cryptography and Security |
| url | https://arxiv.org/abs/2601.21066 |