Research map / Detection and segmentation / CNN features
On the Detection of Digital Face Manipulation
Check 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 supervised direct-regression attention | Xception · DFFD test | TDR at FDR 0.01%: 77.42 -> 83.83 | Helped |
| Add supervised direct-regression attention | Xception · DFFD test | PBCA: 88.44 | Helped |
| Train on DFFD instead of UADFV and add supervised direct-regression attention | Xception · Celeb-DF | AUC: 52.2 -> 71.2 | Helped |
| Add supervised direct-regression attention | Xception trained on DFFD · UADFV | AUC: 75.6 -> 84.2 | Helped |
| Remove attention-map supervision (ablation) | Xception + direct-regression attention · DFFD test | TDR at FDR 0.01%: 83.83 -> 77.07 | Hurt |
| Replace full mask supervision with weak map supervision (ablation) | Xception + direct-regression attention · DFFD test | TDR at FDR 0.01%: 83.83 -> 46.57 | Hurt |
| Use the attention-map average instead of softmax output for detection (ablation) | Xception + supervised direct-regression attention · DFFD test | TDR at FDR 0.01%: 83.83 -> 48.54 | Hurt |
| Replace direct regression with MAM (ablation) | Xception with supervised attention · DFFD test | TDR at FDR 0.01%: 83.83 -> 77.72 | Hurt |
| Use MAM rather than direct regression with weak supervision (ablation) | Xception with attention · DFFD test | TDR at FDR 0.01%: 46.57 -> 72.47 | Helped |
| Add supervised MAM attention | VGG16 · DFFD test | TDR at FDR 0.01%: 0.00 -> 75.89 | Helped |
| Use MAM instead of direct regression (ablation) | VGG16 with supervised attention · DFFD test | TDR at FDR 0.01%: 44.16 -> 75.89 | Helped |
| Move the attention layer later in the network (ablation) | Xception · DFFD test | PBCA: 83.30 -> 88.96 | Helped |
Closest papers by meaning
- Local Relation Learning for Face Forgery Detection (AAAI 2021 · similarity 0.90)
- DevFD : Developmental Face Forgery Detection by Learning Shared and Orthogonal LoRA Subspaces (NeurIPS 2025 · similarity 0.86)
- Forensics Adapter: Adapting CLIP for Generalizable Face Forgery Detection (CVPR 2025 · similarity 0.85)
- Learning Discriminative Noise Guidance for Image Forgery Detection and Localization (AAAI 2024 · similarity 0.85)
- FakeDiffer: Distributional Disparity Learning on Differentiated Reconstruction for Face Forgery Detection (AAAI 2025 · similarity 0.85)
- Rethinking Vision-Language Model in Face Forensics: Multi-Modal Interpretable Forged Face Detector (CVPR 2025 · similarity 0.85)
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.