Research map / Label-efficient and robust learning / Domain adaptation and generalization
Learning Representation and Synergy Invariances: A Provable Framework for Generalized Multimodal Face Anti-Spoofing
Xun Lin, Shuai Wang, Yi Yu, Zitong Yu, Zhou Jiale, Ying Liu, Xiaochun Cao, Alex Chichung Kot, Yefeng Zheng
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 |
|---|---|---|---|
| use RiSe (AsyIRM + MMSD) instead of DADM | multimodal DG FAS · Protocol 1 average | HTER: 13.63 -> 7.70 | Improved |
| use RiSe instead of MMDG++ | multimodal DG FAS · Protocol 1 average | HTER: 9.98 -> 7.70 | Improved |
| use RiSe instead of FLIP (unimodal CLIP) | FAS with CLIP backbone · Protocol 1 average | HTER: 16.11 -> 7.70 | Improved |
| use RiSe instead of MMDG++ when depth or IR is missing at test | multimodal DG FAS · Protocol 2 average | HTER: 16.10 -> 13.52 | Improved |
| use RiSe instead of DADM when depth is missing | multimodal DG FAS · Protocol 2 Missing DEP | HTER: 21.56 -> 13.52 | Improved |
| use RiSe instead of MMDG++ with train-time modality dropout | multimodal DG FAS · Protocol 2+ average | HTER: 15.10 -> 12.10 | Improved |
| use RiSe instead of MMDG on hardest split | multimodal DG FAS · Protocol 2+ PSW->C | HTER: 33.27 -> 20.18 | Improved |
| use RiSe instead of MMDG++ with two source domains | multimodal DG FAS · Protocol 3 CW->PS | HTER: 10.67 -> 8.05 | Improved |
| use RiSe instead of DADM with two source domains | multimodal DG FAS · Protocol 3 PS->CW | HTER: 20.40 -> 19.46 | Improved |
| use RiSe instead of DADM in Protocol 1 AUC | multimodal DG FAS · Protocol 1 average | AUC: 92.96 -> 96.69 | Improved |
Closest papers by meaning
- mmFAS: Multimodal Face Anti-Spoofing Using Multi-Level Alignment and Switch-Attention Fusion (AAAI 2025 · similarity 0.91)
- FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models (AAAI 2026 · similarity 0.89)
- Test-Time Domain Generalization for Face Anti-Spoofing (CVPR 2024 · similarity 0.88)
- Interpretable Face Anti-Spoofing: Enhancing Generalization with Multimodal Large Language Models (AAAI 2025 · similarity 0.88)
- Gradient Alignment for Cross-Domain Face Anti-Spoofing (CVPR 2024 · similarity 0.88)
- From Intuition to Investigation: A Tool-Augmented Reasoning MLLM Framework for Generalizable Face Anti-Spoofing (CVPR 2026 · 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.
