SAAM：针对单目深度估计的隐蔽对抗攻击
SAAM: Stealthy Adversarial Attack on Monocular Depth Estimation
AI总结：
本文提出SAAM，一种针对单目深度估计的隐蔽对抗攻击，通过自然外观补丁造成60%深度误差并影响99%区域，揭示边缘设备上MDE的安全威胁。
摘要：
本文研究了单目深度估计（MDE）对对抗性补丁的脆弱性。我们提出了一种新颖的针对MDE的隐蔽对抗攻击（SAAM），通过破坏估计距离或使物体无缝融入其周围环境来损害MDE。实验表明，所设计的隐蔽补丁能成功导致基于深度神经网络（DNN）的MDE错误估计物体的深度。事实上，我们提出的对抗性补丁在受影响区域比例达到99%的情况下，实现了显著的60%深度误差。重要的是，尽管具有对抗性，该补丁仍保持自然外观，使其对人类观察者不易察觉。我们相信这项工作揭示了边缘设备上MDE面临的对抗性攻击威胁。我们希望它能提高社区对此类攻击潜在现实危害的认识，并鼓励进一步研究开发更稳健和自适应的防御机制。
英文摘要：
In this paper, we investigate the vulnerability of MDE to adversarial patches. We propose a novel \underline{S}tealthy \underline{A}dversarial \underline{A}ttacks on \underline{M}DE (SAAM) that compromises MDE by either corrupting the estimated distance or causing an object to seamlessly blend into its surroundings. Our experiments, demonstrate that the designed stealthy patch successfully causes a DNN-based MDE to misestimate the depth of objects. In fact, our proposed adversarial patch achieves a significant 60\% depth error with 99\% ratio of the affected region. Importantly, despite its adversarial nature, the patch maintains a naturalistic appearance, making it inconspicuous to human observers. We believe that this work sheds light on the threat of adversarial attacks in the context of MDE on edge devices. We hope it raises awareness within the community about the potential real-life harm of such attacks and encourages further research into developing more robust and adaptive defense mechanisms.
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。
