Research map / Detection and segmentation / Open-set and few-shot detection
Fake It till You Make It: Curricular Dynamic Forgery Augmentations towards General Deepfake Detection
Yuzhen Lin, Wentang Song, Bin Li, Yuezun Li, Jiangqun Ni, Han Chen, Qiushi Li
Read the paperCheck 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 |
|---|---|---|---|
| add CDFA training to naive detectors | Xception, EfficientNet-B4, Swin · Cross-dataset average video-level | AUC %: 76.44 -> 91.63 (Swin) | Improved |
| add CDFA training | Xception, EfficientNet-B4, Swin · Cross-manipulation FF++ (trained on DF) | AUC %: FS 53.73 -> 90.04 (Xcep); DF 100 -> 99.54 | Mixed |
| use SSBI instead of BI or SBI as single augmentation (ablation) | CDFA · Cross-dataset ablation | avg AUC %: 83.38 -> 86.67 (BI -> SSBI) | Improved |
| use all augmentations with fixed probability instead of SSBI alone (ablation) | CDFA · Cross-dataset ablation | avg AUC %: 86.67 -> 85.08 | No clear effect |
| add dynamic forgery search (DFS) (ablation) | CDFA with all augmentations · Cross-dataset ablation | avg AUC %: 85.08 -> 88.44 | Improved |
| use o-fake with monotonic curriculum instead of mixing o-fake into p-fake (ablation) | CDFA · Cross-dataset ablation | avg AUC %: 89.32 -> 90.29 | Improved |
| add monotonic curriculum to DFS with o-fake (ablation) | CDFA · Cross-dataset ablation | avg AUC %: 90.97 -> 92.36 | Improved |
| add original fake samples (o-fake) via curriculum (ablation) | BI/SBI/SSBI variants · Cross-dataset ablation | avg AUC %: SBI 82.29 -> 82.92; SSBI 86.67 -> 89.38 | Mixed |
| CDFA versus OST on DFDC frame-level | CDFA · DFDC frame-level | AUC %: 83.30 vs 81.45 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- AUNet — AUNet: Learning Relations Between Action Units for Face Forgery Detection (CVPR 2023) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DF40: Toward Next-Generation Deepfake Detection (NeurIPS 2024 · similarity 0.90)
- Generalizing Deepfake Video Detection with Plug-and-Play: Video-Level Blending and Spatiotemporal Adapter Tuning (CVPR 2025 · similarity 0.90)
- From Specificity to Generality: Revisiting Generalizable Artifacts in Detecting Face Deepfakes (NeurIPS 2025 · similarity 0.89)
- Rethinking Vision-Language Model in Face Forensics: Multi-Modal Interpretable Forged Face Detector (CVPR 2025 · similarity 0.89)
- Exposing the Deception: Uncovering More Forgery Clues for Deepfake Detection (AAAI 2024 · similarity 0.89)
- FreqBlender: Enhancing DeepFake Detection by Blending Frequency Knowledge (NeurIPS 2024 · 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.
