{"slug": "adversarial-tracking-attacks-on-image-sensors-for-object-tracking-in-autonomous", "title": "Adversarial Tracking: Attacks on Image Sensors for Object Tracking in Autonomous Vehicles", "summary": "A paper posted to arXiv (2611.17505v1) reports that adversarial image perturbations sharply degrade object-tracking in autonomous-vehicle perception pipelines, with Higher Order Tracking Accuracy (HOTA) for YOLOv3-based trackers falling from 53.45-56.19% to 14.85-17.03% under image-specific attacks and Multiple Object Tracking Accuracy (MOTA) reaching -772.60% in the worst case. The evaluation, run in the CARLA simulator, paired YOLOv3 and YOLOv4 detectors with the SORT, Deep SORT, Strong SORT and OC-SORT tracking systems; YOLOv4-based trackers proved more robust, with HOTA still dropping to 25.06-26.86%. The authors also show that a single Universal Adversarial Perturbation (UAP) can cut tracking accuracy without a separate perturbation per frame, indicating detection accuracy alone does not characterize downstream trajectory-estimation damage.", "body_md": "arXiv:2611.17505v1 Announce Type: new \nAbstract: Object tracking systems estimate the motion of surrounding vehicles by linking detected objects across video frames. In particular, Tracking-By-Detection (TBD) systems combine object detectors with motion prediction and data association to maintain object identities over time. However, neural network-based detectors are vulnerable to adversarial attacks that perturb input images, and their impact on downstream trajectory estimation is not fully characterized by detection accuracy alone. Therefore, this paper evaluates image-specific and image-agnostic attacks on YOLOv3 and YOLOv4 detectors paired with SORT, Deep SORT, Strong SORT, and OC-SORT tracking systems. The evaluation uses CARLA simulator and measures tracking quality through Multiple Object Tracking Accuracy (MOTA), Higher Order Tracking Accuracy (HOTA), and Identification F1 (IDF1). Under image-specific attacks, HOTA for YOLOv3-based trackers declines from 53.45-56.19% to 14.85-17.03%, while MOTA reaches -772.60% in the worst case. YOLOv4-based trackers exhibit greater robustness, while HOTA still declines to 25.06-26.86%. In addition, image-agnostic attacks can reduce the tracking accuracy using a single Universal Adversarial Perturbation (UAP), without generating a separate perturbation for individual frames.", "url": "https://wpnews.pro/news/adversarial-tracking-attacks-on-image-sensors-for-object-tracking-in-autonomous", "canonical_source": "https://arxiv.org/abs/2611.17505", "published_at": "2026-11-16 05:00:00+00:00", "updated_at": "2026-11-16 05:21:06.310318+00:00", "lang": "en", "topics": ["autonomous-vehicles", "computer-vision", "ai-safety", "ai-research", "machine-learning"], "entities": ["YOLOv3", "YOLOv4", "SORT", "Deep SORT", "Strong SORT", "OC-SORT", "CARLA", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/adversarial-tracking-attacks-on-image-sensors-for-object-tracking-in-autonomous", "markdown": "https://wpnews.pro/news/adversarial-tracking-attacks-on-image-sensors-for-object-tracking-in-autonomous.md", "text": "https://wpnews.pro/news/adversarial-tracking-attacks-on-image-sensors-for-object-tracking-in-autonomous.txt", "jsonld": "https://wpnews.pro/news/adversarial-tracking-attacks-on-image-sensors-for-object-tracking-in-autonomous.jsonld"}}