Alibaba-Tsinghua Adversarial Challenge Dataset

该数据集包含了2017年MSCOCO测试数据中的1000张图片，这些图片曾被用于阿里巴巴-清华对抗性挑战赛中的目标检测任务。此外，该数据集还被用于评估DPAttack方法针对目标检测器的有效性。所有图片的分辨率为500×500，旨在完成目标检测的任务。This dataset contains 1000 images from the 2017 MSCOCO test dataset, which were previously used for the object detection task in the Alibaba-Tsinghua Adversarial Challenge. Furthermore, this dataset has also been utilized to evaluate the effectiveness of the DPAttack method against object detectors. All images have a resolution of 500×500, and the dataset is intended for object detection tasks.