Chengzhi Mao, Mia Chiquier, Hao Wang 0014, Junfeng Yang, Carl Vondrick. Adversarial Attacks are Reversible with Natural Supervision. In 2021 IEEE/CVF International Conference on Computer Vision, ICCV 2021, Montreal, QC, Canada, October 10-17, 2021. pages 641-651, IEEE, 2021. [doi]
Abstract is missing.