Improving the Perturbation-Based Explanation of Deepfake Detectors Through the Use of Adversarially-Generated Samples.

Konstantinos Tsigos, Evlampios Apostolidis, Vasileios Mezaris

Browse the full WACV paper archive.