AntiStyler：利用风格移除防御目标检测模型对抗对抗补丁攻击
AntiStyler: Defending Object Detection Models Against Adversarial Patch Attacks Using Style Removal
提供了一种无需训练、模型无关的对抗补丁防御前处理：通过识别具有攻击性“随机”风格的像素并生成掩码，再用空间滤波增强掩码，可直接插到任意实时目标检测器前，兼顾干净图像性能与10-12 FPS的推理速度。
原文依据 AntiStyler identifies and masks pixels that exhibit a "random" style associated with adversarial attacks
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论对抗补丁攻击对目标检测（OD）模型的可靠性构成重大威胁，在实时安防应用中尤为突出。尽管已有多种防御方法被提出，但它们往往存在两点局限：1）在良性图像上性能下降；2）处理时间过长，难以满足实时目标检测应用的需求。本文提出 AntiStyler，一种新颖且快速的对抗补丁防御方法。给定输入图像，AntiStyler 识别并掩蔽呈现与对抗攻击相关的“随机”风格的像素，并使用一系列空间滤波器增强掩码、去除多余噪声，从而高效地掩蔽对抗补丁。AntiStyler 具备模型无关、补丁无关和攻击无关的特性，且无需任何训练，因此是一种完全无关的零样本对抗补丁攻击防御方法。我们在 COCO、INRIA、Superstore 和 APRICOT 数据集上，针对数字攻击和物理攻击进行了评估，结果表明 AntiStyler 具有最优的鲁棒性（对抗性能提升 8-15 mAP%），同时不损害在良性图像上的原始性能。此外，与大多数现有防御方法不同，AntiStyler 能够以每秒 10-12 帧（FPS）的速度处理，使其高效且适用于实时目标检测应用。
英文原文摘要 1313 字符
Adversarial patch attacks pose a significant threat to the reliability of object detection (OD) models, particularly in real-time security applications. Although several defenses have been proposed, they often suffer from two limitations: 1) reduced performance on benign images, and 2) impractical processing time for real-time OD applications. In this paper, we present AntiStyler, a novel and rapid defense against adversarial patches. Given an input image, AntiStyler identifies and masks pixels that exhibit a "random" style associated with adversarial attacks and uses a series of spatial filters to enhance the mask and remove unwanted noise, efficiently masking adversarial patches. AntiStyler features model-, patch-, and attack-agnostic capabilities and does not require any training, making it a fully agnostic zero-shot defense against adversarial patch attacks. Our evaluation on the COCO, INRIA, Superstore, and APRICOT datasets, with both digital and physical attacks, demonstrates AntiStyler's state-of-the-art robustness (improving adversarial performance by 8-15 mAP%) without compromising the original performance on benign images. Additionally, unlike most existing defenses, AntiStyler can process 10-12 frames per second (FPS), making it efficient and relevant for real-time OD applications.