Potential auto-driving threat: Universal rain-removal attack

Jincheng Hu; Jihao Li; Zhuoran Hou; Jingjing Jiang; Cunjia Liu; Liang Chu; Yanjun Huang; Yuanjian Zhang

doi:10.1016/j.isci.2023.107393

Potential auto-driving threat: Universal rain-removal attack

iScience. 2023 Jul 13;26(9):107393. doi: 10.1016/j.isci.2023.107393. eCollection 2023 Sep 15.

Authors

Jincheng Hu¹, Jihao Li¹, Zhuoran Hou¹, Jingjing Jiang¹, Cunjia Liu¹, Liang Chu², Yanjun Huang³, Yuanjian Zhang¹

Affiliations

¹ Department of Aeronautical and Automotive Engineering, Loughborough University, Leicestershire LE11 3TU UK.
² College of Automotive Engineering, Jilin University, Changchun 130022 China.
³ School of Automotive Studies, Tongji University, Shanghai 201804 China.

Abstract

Severe weather conditions pose a significant challenge for computer vision algorithms in autonomous driving applications, particularly regarding robustness. Image rain-removal algorithms have emerged as a potential solution by leveraging the power of neural networks to restore rain-free backgrounds in images. However, existing research overlooks the vulnerability concerns in neural networks, which exposes a potential threat to the intelligent perception of autonomous vehicles in rainy conditions. This paper proposes a universal rain-removal attack (URA) that exploits the vulnerability of image rain-removal algorithms. By generating a non-additive spatial perturbation, URA significantly diminishes scene restoration similarity and image quality. The imperceptible and generic perturbation employed by URA makes it a crucial tool for vulnerability detection in image rain-removal algorithms and a potential real-world AI attack method. Experimental results demonstrate that URA can reduce scene repair capability by 39.5% and image generation quality by 26.4%, effectively targeting state-of-the-art rain-removal algorithms.

Keywords: Artificial intelligence; Artificial intelligence applications; Computer science.