Face-Feature Tuning is a label-free logit remapping method that reduces FPR/TPR gaps across groups in deepfake detection while preserving overall accuracy.
DF40: Toward next-generation deepfake detection
10 Pith papers cite this work, alongside 2 external citations. Polarity classification is still indexing.
10
Pith papers citing it
2
external citations · external index
verdicts
UNVERDICTED 10representative citing papers
SynCred-Bench shows that 15 MLLMs reach only 10.5% TPR, open-source detectors under 5%, commercial APIs 57.6%, and humans 63% TPR at 5% FPR when identifying AI-generated images with synthetic credibility.
A dual-encoder deepfake detector pairs a frozen specialist with a LoRA-tuned MLLM, trained first via binary alignment then via RL to reward explain-then-classify behavior, yielding improved cross-dataset performance and interpretability.
A replay method for continual face forgery detection condenses real-fake distribution discrepancies into compact maps and synthesizes compatible samples from current real faces to reduce forgetting under tight memory budgets without storing historical images.
DeepSpeak provides over 100 hours of consented, identity-matched real and modern deepfake audiovisual content focused on talking heads, with evaluations showing existing detectors fail to generalize without retraining.
A frequency-aware triple-branch network with mutual information-based decoupling and fusion losses achieves state-of-the-art deepfake detection across six benchmarks.
DLED reformulates open-set face forgery detection as an uncertainty estimation task and uses dual-level spatial-frequency evidence collection to identify novel fake categories, claiming 20% average gains over baselines.
Orthogonal subspace decomposition via SVD on vision foundation model features preserves high-rank pre-trained knowledge by freezing principal components and adapting residuals, reducing overfitting for better generalization in AI-generated image detection.
3D CNN detector with temporal consistency regularizer reaches 92.8% accuracy on DeepfakeTIMIT and 76.4% cross-dataset on FaceForensics++ without fine-tuning.
The NTIRE 2026 challenge finds that large foundation models combined with ensembles and degradation-aware training produce the most robust deepfake detectors.
citing papers explorer
-
Toward Calibrated, Fair, and accurate Deepfake Detection
Face-Feature Tuning is a label-free logit remapping method that reduces FPR/TPR gaps across groups in deepfake detection while preserving overall accuracy.
-
SynCred-Bench: Benchmarking Synthetic Credibility in AI-Generated Visual Misinformation
SynCred-Bench shows that 15 MLLMs reach only 10.5% TPR, open-source detectors under 5%, commercial APIs 57.6%, and humans 63% TPR at 5% FPR when identifying AI-generated images with synthetic credibility.
-
The Regularizing Power of Language-Training Deepfake Detectors
A dual-encoder deepfake detector pairs a frozen specialist with a LoRA-tuned MLLM, trained first via binary alignment then via RL to reward explain-then-classify behavior, yielding improved cross-dataset performance and interpretability.
-
Direct Discrepancy Replay: Distribution-Discrepancy Condensation and Manifold-Consistent Replay for Continual Face Forgery Detection
A replay method for continual face forgery detection condenses real-fake distribution discrepancies into compact maps and synthesizes compatible samples from current real faces to reduce forgetting under tight memory budgets without storing historical images.
-
The DeepSpeak Dataset
DeepSpeak provides over 100 hours of consented, identity-matched real and modern deepfake audiovisual content focused on talking heads, with evaluations showing existing detectors fail to generalize without retraining.
-
Unveiling Deepfakes: A Frequency-Aware Triple Branch Network for Deepfake Detection
A frequency-aware triple-branch network with mutual information-based decoupling and fusion losses achieves state-of-the-art deepfake detection across six benchmarks.
-
Open Set Face Forgery Detection via Dual-Level Evidence Collection
DLED reformulates open-set face forgery detection as an uncertainty estimation task and uses dual-level spatial-frequency evidence collection to identify novel fake categories, claiming 20% average gains over baselines.
-
Orthogonal Subspace Decomposition for Generalizable AI-Generated Image Detection
Orthogonal subspace decomposition via SVD on vision foundation model features preserves high-rank pre-trained knowledge by freezing principal components and adapting residuals, reducing overfitting for better generalization in AI-generated image detection.
-
Deepfake Detection in Social Media: A Temporal Artifact Analysis Using 3D Convolutional Neural Networks
3D CNN detector with temporal consistency regularizer reaches 92.8% accuracy on DeepfakeTIMIT and 76.4% cross-dataset on FaceForensics++ without fine-tuning.
-
Robust Deepfake Detection, NTIRE 2026 Challenge: Report
The NTIRE 2026 challenge finds that large foundation models combined with ensembles and degradation-aware training produce the most robust deepfake detectors.