A Survey on Adversarial Attack in the Age of Artificial Intelligence

Table 6

Common datasets for image adversarial attack.

Type of dataset	Data source	Application instances
Publicly accessible dataset	ImageNet	Xiao et al. 2019 [55]; Ma et al. 2019 [92]
	MNIST	Demontis et al. 2019 [20]; Ling et al. 2019 [93]; Fang et al. 2020 [94]; Ma et al. 2019 [92]; Moosavi-Dezfooli et al. 2016 [36]; Yang et al. 2019 [95]
	CIFAR-10	Ling et al. 2019 [93]; Shafahi et al. 2018 [52]; Ma et al. 2019 [92]; Moosavi-Dezfooli et al. 2016 [36]; Yang et al. 2019 [95]
	CH-MNIST; Fashion-MNIST; Breast Cancer Wisconsin	Fang et al. 2020 [94]
	VidTIMIT database	Korshunov et al. 2018 [96]
	WebFace; VGGFace2	Shan et al. 2019 [97]
	FaceScrub	Yang et al.2019 [95]; Shan et al. 2019 [97]
	PubFig	Sharif et al. 2016 [50]; Shan et al. 2019 [97]
	Cora; Citeseer; Polblogs	Jin et al. 2020 [98]
	Social Face Classification (SFC) dataset	Taigman et al. 2014 [99]
	MS-COCO	Chen et al. 2019 [54]
	CelebA	Yang et al. 2019 [95]
	MS-COCO 2017; PASCAL VOC 2007; PASCAL VOC 2012	Wang et al. 2020 [56]
	Labeled Faces in the Wild (LFW) database	Demontis et al. 2019 [20]; Taigman et al. 2014 [99]; Ma et al. 2019 [92]
	YouTube Faces (YTF) dataset	Taigman et al. 2014 [99]
	LIDC-IDRI dataset	Mirsky et al. 2019 [53]
	ILSVRC 2012	Simonyan et al.2015 [100]; Moosavi-Dezfooli et al. 2016 [36]
Commercial dataset	Fugazi	Din et al. 2018 [101]
Artificially generated dataset	Generated by toolkits manually	Yu et al. 2020 [102]