Research map / Label-efficient and robust learning / Domain adaptation and generalization
Generalizable Representation Learning for Mixture Domain Face Anti-Spoofing
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 |
|---|---|---|---|
| Iteratively cluster unlabeled mixture domains and train with meta-learning | D²AM · O&C&M to I | HTER (%): 17.30 -> 15.43 | Helped |
| Iteratively cluster unlabeled mixture domains and train with meta-learning | D²AM · O&C&I to M | HTER (%): 13.89 -> 12.70 | Helped |
| Iteratively cluster unlabeled mixture domains and train with meta-learning | D²AM · O&M&I to C | HTER (%): 20.27 -> 20.98 | Hurt |
| Iteratively cluster unlabeled mixture domains and train with meta-learning | D²AM · I&C&M to O | HTER (%): 16.45 -> 15.27 | Helped |
| Iteratively cluster unlabeled mixture domains and train with meta-learning | D²AM · M&I to C | AUC (%): 71.29 -> 72.04 | Helped |
| Iteratively cluster unlabeled mixture domains and train with meta-learning | D²AM · M&I to O | HTER (%): 29.12 -> 27.70 | Helped |
| Use provided domain labels instead of the method's pseudo-domain assignment (ablation) | D²AM · O&C&M to I | HTER (%): 15.43 -> 18.24 | Hurt |
| Randomly select meta-train and meta-test domains instead of using assigned domains (ablation) | D²AM · O&C&M to I | HTER (%): 15.43 -> 27.05 | Hurt |
| Use all channels rather than low-attention channels for clustering (ablation) | D²AM · O&C&M to I | HTER (%): 15.43 -> 16.57 | Hurt |
| Add domain enhancement entropy loss to suppress task information in domain features (ablation) | D²AM · O&C&M to I | HTER (%): 15.89 -> 15.43 | Helped |
| Add MMD-based regularization (ablation) | D²AM · O&C&M to I | HTER (%): 16.11 -> 15.43 | Helped |
| Set the number of clusters to 2 rather than 3 (ablation) | D²AM · O&C&M to I | HTER (%): 15.43 -> 17.16 | Hurt |
Closest papers by meaning
- Test-Time Domain Generalization for Face Anti-Spoofing (CVPR 2024 · similarity 0.91)
- Gradient Alignment for Cross-Domain Face Anti-Spoofing (CVPR 2024 · similarity 0.91)
- Rethinking Generalizable Face Anti-spoofing via Hierarchical Prototype-guided Distribution Refinement in Hyperbolic Space (CVPR 2024 · similarity 0.89)
- Domain-Hallucinated Updating for Multi-Domain Face Anti-spoofing (AAAI 2024 · similarity 0.89)
- CFPL-FAS: Class Free Prompt Learning for Generalizable Face Anti-spoofing (CVPR 2024 · similarity 0.89)
- FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models (AAAI 2026 · similarity 0.87)
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.