arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2402.01220cs.CVcs.CR
收藏

深入研究语义分割上基于决策的黑盒攻击

Delving into Decision-based Black-box Attacks on Semantic Segmentation

Shanghai Engineering Research Center of AI & Robotics(上海市人工智能与机器人工程研究中心);Academy for Engineering & Technology(工程技术学院);Fudan University(复旦大学);Shanghai Key Lab of Intelligent Information Processing(上海市智能信息处理重点实验室);School of Computer Science, Fudan University(复旦大学计算机科学技术学院)

Zhaoyu Chen, Zhengyang Shan, Jingwen Chang, Kaixun Jiang, Dingkang Yang, Yiting Cheng, Wenqiang Zhang

AI总结:

本文首次探索语义分割的基于决策的黑盒攻击,提出离散线性攻击(DLA)方法,通过离散线性噪声实现高效攻击,实验证明其仅用50次查询即可大幅降低PSPNet的mIoU。

摘要:

语义分割是一项基础视觉任务,在涉及安全敏感考量的应用中得到广泛部署。然而,近期研究表明语义分割模型易受白盒攻击的对抗性影响,但其针对黑盒攻击的对抗鲁棒性尚未得到充分探究。本文首次对语义分割的基于决策的黑盒攻击展开探索。首先,我们通过案例研究分析了语义分割给基于决策的攻击带来的挑战。随后,为应对这些挑战,我们提出了一种针对语义分割的基于决策的攻击方法,名为离散线性攻击(Discrete Linear Attack,DLA)。该方法基于随机搜索和代理指标,利用离散线性噪声进行扰动探索与校准,以实现高效的攻击效率。我们在Cityscapes和ADE20K两个数据集的5个模型上,针对8种攻击开展了对抗鲁棒性评估。DLA在Cityscapes上展现出强大威力:仅需50次查询,就将PSPNet的mIoU从可观的77.83%大幅降至仅2.14%。

英文摘要:

Semantic segmentation is a fundamental visual task that finds extensive deployment in applications with security-sensitive considerations. Nonetheless, recent work illustrates the adversarial vulnerability of semantic segmentation models to white-box attacks. However, its adversarial robustness against black-box attacks has not been fully explored. In this paper, we present the first exploration of black-box decision-based attacks on semantic segmentation. First, we analyze the challenges that semantic segmentation brings to decision-based attacks through the case study. Then, to address these challenges, we first propose a decision-based attack on semantic segmentation, called Discrete Linear Attack (DLA). Based on random search and proxy index, we utilize the discrete linear noises for perturbation exploration and calibration to achieve efficient attack efficiency. We conduct adversarial robustness evaluation on 5 models from Cityscapes and ADE20K under 8 attacks. DLA shows its formidable power on Cityscapes by dramatically reducing PSPNet's mIoU from an impressive 77.83% to a mere 2.14% with just 50 queries.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑