Complete Defense Framework to Protect Deep Neural Networks against Adversarial Examples

Table 2

The classification accuracy of adversarial examples generated by evaluated attacks on the testing set (%).

Attacks	FGSM	R-FGSM	BIM	UAP	DeepFool	CW_UT	CW_T
Accuracy	26.1	12.8	0	3.5	0.2	0.7	1.0