We study black-box adversarial attacks for image classifiers in a constrained threat model, where adversaries can only modify a small fraction of pixels in the form of scratches on an image. We show that it is possible for adversaries to generate localized textit{adversarial scratches} that cover less than $5%$ of the pixels in an image and achieve targeted success rates of $98.77%$ and $97.20%$ on ImageNet and CIFAR-10 trained ResNet-50 models, respectively. We demonstrate that our scratches are effective under diverse shapes, such as straight lines or parabolic Baezier curves, with single or multiple colors. In an extreme condition, in which our scratches are a single color, we obtain a targeted attack success rate of $66%$ on CIFAR-10 with an order of magnitude fewer queries than comparable attacks. We successfully launch our attack against Microsofts Cognitive Services Image Captioning API and propose various mitigation strategies.
تحميل البحث