A comprehensive evaluation framework and efficient regularization strategy for adversarial robustness of deepfake detectors under closed and open set scenarios

Published in Computer Vision and Image Understanding, 2026

Direct Link