Curvelet-based frequency enhancement with attention mechanisms achieves 98.48% accuracy and 99.96% AUC on low-compression FaceForensics++ deepfakes while performing well under high compression.
Title resolution pending
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
-
Curvelet-Based Frequency-Aware Feature Enhancement for Deepfake Detection
Curvelet-based frequency enhancement with attention mechanisms achieves 98.48% accuracy and 99.96% AUC on low-compression FaceForensics++ deepfakes while performing well under high compression.