Adversarial Attacks for Salient Object Detection

This dataset is an image repository that contains five different image databases to evaluate adversarial robustness in Salient Object Detection (SOD) through the introduction of 12 adversarial examples, each leveraging a known adversarial attack or noise perturbation. The dataset comprises 56,387 digital images, resulting from applying adversarial examples on subsets of four standard databases (i.e. FT, PASCAL-S, ImgSal, DUTS), and a fifth database (SNPL) portraying a real-world visual attention problem of a shorebird called the Snowy Plover. Original and rescaled images from the five databases used with the adversarial examples are also included as part of this dataset for ease of access and explainability.本数据集为图像库，包含五个独立图像数据库，用于通过引入12组对抗样本评估显著目标检测（Salient Object Detection, SOD）的对抗鲁棒性，每组对抗样本均采用一种已公开的对抗攻击或噪声扰动方法生成。该数据集共计包含56387张数字图像，由两部分构成：一是对四个标准数据库（即FT、PASCAL-S、ImgSal、DUTS）的子集施加对抗样本得到的图像，二是第五个数据库SNPL的相关数据——该数据库描绘了名为雪鸻（Snowy Plover）的滨鸟的真实视觉注意力任务场景。为便于访问与可解释性分析，本数据集同时收录了与上述对抗样本配套的五个数据库的原始图像与缩放后图像。