Research map / Label-efficient and robust learning / Domain adaptation and generalization
Towards Data-Centric Face Anti-spoofing: Improving Cross-Domain Generalization via Physics-Based Data Synthesis
Rizhao Cai, Cecelia Soh, Zitong Yu, Haoliang Li, Wenhan Yang, Alex Chichung Kot
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 |
|---|---|---|---|
| replace traditional augmentation with FAS-Aug | ViT-Convpass · MICO average | HTER (%): 11.53 -> 8.93 | Improved |
| add Con loss and SARE to FAS-Aug | ViT-Convpass · MICO average | HTER (%): 8.93 -> 7.39 | Improved |
| replace traditional augmentation with FAS-Aug | ResNet-18 · O&M&I to C | HTER (%): 18.00 -> 21.50 | Hurt |
| add supervised contrastive loss on real faces (ablation) | ViT-Convpass · I&C&M to O | HTER (%): 8.31 -> 7.43 | Improved |
| add SARE loss (ablation) | ViT-Convpass · I&C&M to O | HTER (%): 8.31 -> 7.81 | Improved |
| add both Con loss and SARE (ablation) | ViT-Convpass · I&C&M to O | HTER (%): 8.31 -> 6.77 | Improved |
| add extra real-face data (SiW, ROSE-YOUTU) with traditional augmentation | ViT-Convpass&TA · C&I&O to M | HTER (%): 13.33 -> 24.03 | Hurt |
| add FAS-Aug, Con and SARE when extra real faces are used | ViT-Convpass · C&I&O+SY to M | HTER (%): 24.03 -> 6.92 | Improved |
| replace traditional augmentation with FAS-Aug | SSDG-R* · C&M&I to O | HTER (%): 16.73 -> 15.25 | Improved |
| replace traditional augmentation with FAS-Aug | SSAN-R* · C&M&I to O | HTER (%): 16.92 -> 13.79 | Improved |
| replace traditional augmentation with FAS-Aug | FLIP-MCL* · C&M&I to O | HTER (%): 8.69 -> 7.41 | Improved |
| combine FAS-Aug with traditional augmentation | ViT-ConvPass · I&C&M to O | HTER (%): 10.74 -> 6.71 | Improved |
Closest papers by meaning
- Test-Time Domain Generalization for Face Anti-Spoofing (CVPR 2024 · similarity 0.92)
- Gradient Alignment for Cross-Domain Face Anti-Spoofing (CVPR 2024 · similarity 0.91)
- From Intuition to Investigation: A Tool-Augmented Reasoning MLLM Framework for Generalizable Face Anti-Spoofing (CVPR 2026 · similarity 0.90)
- FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models (AAAI 2026 · similarity 0.89)
- Rethinking Generalizable Face Anti-spoofing via Hierarchical Prototype-guided Distribution Refinement in Hyperbolic Space (CVPR 2024 · similarity 0.89)
- mmFAS: Multimodal Face Anti-Spoofing Using Multi-Level Alignment and Switch-Attention Fusion (AAAI 2025 · similarity 0.89)
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.
