Research map / Video / Video segmentation and tracking
DeeperForensics-1.0: A Large-Scale Dataset for Real-World Face Forgery Detection
Liming Jiang, Ren Li, Wayne Wu, Chen Qian, Chen Change Loy
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Train on DeeperForensics-1.0 standard data instead of FF++ DeepFakes data | I3D · Hidden test set | accuracy: 56.63 -> 79.25 | Improved |
| Add randomly mixed distortions to the standard training set (ablation) | I3D · Hidden test set | accuracy: 79.25 -> 80.13 | Improved |
| Add single-level distortions to the standard training set (ablation) | I3D · Hidden test set | accuracy: 79.25 -> 80.13 | Improved |
| Add random-level distortions to the standard training set (ablation) | I3D · Hidden test set | accuracy: 79.25 -> 80.13 | Improved |
| Test on single-level distorted data instead of clean standard data (ablation) | I3D · DeeperForensics-1.0 | accuracy: 100.00 -> 90.75 | Hurt |
| Test on random-level distorted data instead of clean standard data (ablation) | I3D · DeeperForensics-1.0 | accuracy: 100.00 -> 96.88 | Hurt |
| Add MAdaIN style matching and fusion (ablation) | DF-VAE · Many-to-many face swapping | visual fidelity | Improved |
| Add hourglass structure extraction (ablation) | DF-VAE · Many-to-many face swapping | disentanglement quality | Improved |
| Construct unpaired data in the same domain (ablation) | DF-VAE · Many-to-many face swapping | visual fidelity | Improved |
| Add temporal consistency loss (ablation) | DF-VAE · Self-reenactment | per-pixel re-rendering error | Improved |
| Use DeeperForensics-1.0 generated fake videos | DF-VAE dataset · Dataset user study | real rating: 61.0% -> 64.1% | Improved |
| Use DF-VAE rather than the best listed baseline, faceswap-GAN | Face swapping methods · Face swapping evaluation | FID: 24.718 -> 22.097 | Improved |
Closest papers by meaning
- FaceForensics++: Learning to Detect Manipulated Facial Images (ICCV 2019 · similarity 0.93)
- DF40: Toward Next-Generation Deepfake Detection (NeurIPS 2024 · similarity 0.89)
- Face Forgery Video Detection via Temporal Forgery Cue Unraveling (CVPR 2025 · similarity 0.88)
- Identity-Aware Vision-Language Model for Explainable Face Forgery Detection (AAAI 2026 · similarity 0.88)
- Forensics Adapter: Adapting CLIP for Generalizable Face Forgery Detection (CVPR 2025 · similarity 0.88)
- Generalizing Deepfake Video Detection with Plug-and-Play: Video-Level Blending and Spatiotemporal Adapter Tuning (CVPR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
