AdvLIM-LD: Toward Cross-Task Physical Adversarial Attack via LED Illumination Modulation for Lane Detection Systems
AdvLIM-LD: кросс-задаковая физическая адверсариальная атака на системы обнаружения дорожной разметки с помощью модуляции освещения светодиодами
2025-08-18
SCID: 54.1/7p4vnc4e
Discuss with AI
LED illumination modulationcross-task transferabilitylane detection systemsphysical adversarial attackrolling shutter effect
Figures from the paper
Abstract (AI)
The reliability and robustness of lane detection play an instrumental role in the practical deployment of autonomous driving systems. Despite previous research indicating that adversarial examples can negatively affect lane detection models, leading to erroneous lane detection, most existing adversarial attacks are designed to attack lane detection models of a single deep learning-based task. To address this issue, we propose a towards cross-task physical adversarial attack via LED illumination modulation for lane detection systems (AdvLIM-LD) by analyzing the common critical feature response mechanisms of lane detection models of different task types. The method first introduces a CFD module to evaluate the impact of adversarial examples on critical features contributing to various tasks. Then, it utilizes the pulse width modulation of LEDs and the rolling shutter effect of CMOS image sensor to implant imperceptible adversarial perturbations during the image acquisition process. Finally, it enhances the cross-task transferability of adversarial examples by jointly optimizing a dual objective function composed of critical feature loss and model output accuracy loss. In digital domain simulations, adversarial examples generated by AdvLIM-LD caused an average detection accuracy degradation of 74.64% across lane detection models representing different task types, including the segmentation model SCNN, the anchor-based detector LaneATT, the curve-fitting model LSTR, and the keypoint detection model GANet. For cross-task attacks, the average accuracy degradation reached 55.37%. Physical-world experiments further validated the effectiveness of AdvLIM-LD, demonstrating its ability to degrade the source models’ average detection accuracy by over 80%.
Key Findings
1
A CFD module evaluates adversarial effects on critical features shared across lane detection models with different task formulations.
2
AdvLIM-LD introduces a cross-task physical adversarial attack against lane detection systems using LED illumination modulation.
3
Cross-task attacks achieved 55.37% average accuracy degradation, while physical experiments reduced source-model accuracy by over 80%.
4
Digital simulations produced an average 74.64% accuracy degradation across SCNN, LaneATT, LSTR, and GANet lane detection models.
5
LED pulse-width modulation combined with CMOS rolling-shutter effects implants imperceptible perturbations during image acquisition.
Research Object
Lane detection systems and their deep learning models under LED-based physical adversarial illumination
Research Subject
Cross-task vulnerability and detection-accuracy degradation caused by imperceptible LED-modulated adversarial perturbations across segmentation, anchor-based detection, curve-fitting, and keypoint detection models
Publication Details
Publication Date
2025-08-18
Journal
Publisher
ISSN
Cited by
1
Access Type
Author Information
Download PDF
Subscribe to digest