M3D-Net reconstructs 3D facial features from RGB images and fuses them with RGB features through attention-based modules to achieve claimed state-of-the-art deepfake detection.
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
-
M3D-Net: Multi-Modal 3D Facial Feature Reconstruction Network for Deepfake Detection
M3D-Net reconstructs 3D facial features from RGB images and fuses them with RGB features through attention-based modules to achieve claimed state-of-the-art deepfake detection.