Research map / Label-efficient and robust learning / OOD and anomaly detection
Implicit Identity Leakage: The Stumbling Block to Improving Deepfake Detection Generalization
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 data augmentations to the baseline (ablation) | ResNet-34 deepfake detector · Celeb-DF | AUC (%): 64.05 -> 80.07 | Helped |
| Add data augmentations to the baseline (ablation) | ResNet-34 deepfake detector · DFDC-V2 | AUC (%): 48.73 -> 62.46 | Helped |
| Add MFS without data augmentations or ADM (ablation) | ResNet-34 deepfake detector · Celeb-DF | AUC (%): 64.05 -> 76.35 | Helped |
| Add MFS without data augmentations or ADM (ablation) | ResNet-34 deepfake detector · DFDC-V2 | AUC (%): 48.73 -> 59.97 | Helped |
| Add MFS to data augmentations (ablation) | ResNet-34 deepfake detector · Celeb-DF | AUC (%): 80.07 -> 86.68 | Helped |
| Add MFS to data augmentations (ablation) | ResNet-34 deepfake detector · DFDC-V2 | AUC (%): 62.46 -> 67.94 | Helped |
| Add ADM to data augmentations and MFS (ablation) | ResNet-34 deepfake detector · Celeb-DF | AUC (%): 86.68 -> 91.15 | Helped |
| Add ADM to data augmentations and MFS (ablation) | ResNet-34 deepfake detector · DFDC-V2 | AUC (%): 67.94 -> 71.49 | Helped |
| Add data augmentations, MFS, and ADM (ablation) | ResNet-34 deepfake detector · FF++ | AUC (%): 99.88 -> 99.70 | Hurt |
| Add the ID-unaware method to SBI (ablation) | SBI · FF++ | AUC (%): 99.64 -> 99.33 (↓0.31) | Hurt |
| Add the ID-unaware method to SBI (ablation) | SBI · Celeb-DF | AUC (%): 93.18 -> 94.15 (↑ 0.97) | Helped |
| Add the ID-unaware method to SBI (ablation) | SBI · DFDC-V2 | AUC (%): 72.42 -> 79.57 (↑ 7.15) | Helped |
Closest papers by meaning
- D^3: Scaling Up Deepfake Detection by Learning from Discrepancy (CVPR 2025 · similarity 0.88)
- Exposing the Deception: Uncovering More Forgery Clues for Deepfake Detection (AAAI 2024 · similarity 0.88)
- DiffusionFake: Enhancing Generalization in Deepfake Detection via Guided Stable Diffusion (NeurIPS 2024 · similarity 0.87)
- Exploring Unbiased Deepfake Detection via Token-Level Shuffling and Mixing (AAAI 2025 · similarity 0.87)
- From Specificity to Generality: Revisiting Generalizable Artifacts in Detecting Face Deepfakes (NeurIPS 2025 · similarity 0.87)
- Preserving Fairness Generalization in Deepfake Detection (CVPR 2024 · similarity 0.86)
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.