With the development of applications like autonomous driving, object detection has gained significant attention, while also highlighting critical vulnerabilities like backdoor attacks that severely co...
The paper introduces TRIM, a black‑box defense for backdoor attacks in computer vision models. TRIM identifies and removes malicious trigger regions at inference time using region‑based segmentation, adaptive trigger discovery via inpainting and diffusion, and selective purification, without needing model internals, training data, or clean samples. Experiments on various datasets and trigger types show TRIM reduces attack success rates to as low as 1.16% while maintaining high clean accuracy.
By Ahmed Abdelnaby, Mohamed Elmahallawy
arXiv:2607. 06592v1 Announce Type: cross Abstract: Object detectors have many applications in safety-critical systems, but they are known to be sensitive to worst-case perturbations such as adversarial attacks, which limits their applicability in real-world scenarios.
By Vincent L\'eb\'e (IRIT, DTIPG - SNCF, UT3), Yannick Prudent (IRIT, DTIPG - SNCF, UT3), Corentin Friedrich (IRIT, DTIPG - SNCF, UT3), Thomas Massena (IRIT, DTIPG - SNCF, UT3), Ronan Sicre (IRIT), Franck Mamalet
Object detectors have many applications in safety-critical systems, but they are known to be sensitive to worst-case perturbations such as adversarial attacks, which limits their applicability in real-world scenarios. Compared with classification, adversarial robustness for object detection has received less attention, and existing methods are often tied to adversarial training, whose performance may not transfer across attacks, perturbation budgets, or architectures.
Camera-based object detectors are vulnerable to physical adversarial attacks designed to suppress detections. While adversarial training and input purification offer some protection, they often overfit to specific attack distributions and fail on adaptive adversaries.
The paper introduces DistScan, a backdoor detection framework for object detection models that identifies malicious behavior by detecting shifts in the pre‑NMS prediction class distribution relative to training class frequencies. DistScan operates on clean validation data, requiring no access to model weights, trigger knowledge, or additional training, and it aggregates intermediate predictions to flag backdoored models. Experiments on MS‑COCO and PASCAL VOC across two architectures and three scene‑level attack scenarios show that DistScan outperforms existing methods, improving average detection accuracy by 27.32 percentage points over the best baseline.
By Longtian Wang, Zhengyu Zhao, Chenhao Lin, Le Yang, Shiwei Wang, Yuhan Zhi, Xiaofei Xie, Chao Shen
arXiv:2608. 16031v1 Announce Type: new Abstract: Camera-based object detectors are vulnerable to physical adversarial attacks designed to suppress detections.
By Yuting Wu, Dongfang Guo, Xiangzhong Luo, Qun Song, Rui Tan
RePAIR is a post‑hoc, label‑free defense for semantic segmentation models that counters backdoor attacks by inferring victim‑target relationships from unlabeled images and correcting predictions in feature space and output space. It localizes suspicious pixels by comparing classifier head outputs with internal semantic evidence, applying conservative constraints to avoid harming clean predictions. Experiments demonstrate that RePAIR significantly reduces attack success across various segmentation architectures while largely preserving clean performance.
By Arnold Brosch, Kira Maag
Vision foundation models such as DINOv2 enable strong few‑shot anomaly detection (FSAD) using a frozen backbone and a non‑parametric k‑nearest‑neighbor (k‑NN) scorer. The paper introduces a white‑box attack framework for k‑NN memory‑bank detectors and proposes an adversarially robust FSAD method that is training‑free with respect to the backbone. The defense combines a PatchShift input augmentation and a FeaturePurifier residual MLP, achieving significant gains in image‑level AUROC (~27%) and pixel‑level PRO (~51%) on MVTec‑AD, VisA, and MPDD while maintaining clean accuracy within 3%.
whyItMatters:"The work addresses the previously unstudied few‑shot regime for robust anomaly detection, providing a practical, training‑free defense that matches full‑shot adversarially‑trained baselines and improves performance under adversarial attacks."
By Akib Mohammed Khan, Bartosz Krawczyk
The paper introduces Poison-then-Hide, a finetuning-activated backdoor attack that implants a dormant backdoor into a pretrained vision encoder and activates it during full-model finetuning on clean data. The method combines trigger optimization, base encoder poisoning, and targeted unlearning to conceal the backdoor, achieving up to 97% success across six datasets and three architectures. The study shows that standard detection and mitigation defenses are ineffective, as the backdoor can reemerge after benign finetuning.
arXiv:2609.31558v1 Announce Type: new
Abstract: Contrastive Language--Image Pretraining (CLIP) has emerged as a dominant vision backbone due to its strong transferability and zero-shot capabilities....
By Ahmed Abdelnaby, Mohamed Elmahallawy
arXiv:2607. 05516v2 Announce Type: replace-cross Abstract: Model-specific adversarial attacks have been extensively studied.
By Paul K. Mandal, Pavan Reddy, Tristan Malatynski