A multi-stream ensemble using DINOv2 and CLIP backbones trained with extreme degradations achieves stable deepfake detection and fourth place in the NTIRE 2026 challenge.
Practical manipulation model for robust deepfake detection, 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Robust Deepfake Detection: Mitigating Spatial Attention Drift via Calibrated Complementary Ensembles
A multi-stream ensemble using DINOv2 and CLIP backbones trained with extreme degradations achieves stable deepfake detection and fourth place in the NTIRE 2026 challenge.