Research map / Label-efficient and robust learning / Domain adaptation and generalization
Learning Unknown Spoof Prompts for Generalized Face Anti-Spoofing Using Only Real Face Images
Fangling Jiang, Qi Li, Weining Wang, Wei Shen, Bing Liu, Zhenan Sun
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 |
|---|---|---|---|
| add one-class discriminative classification loss l_dc (ablation) | Ours · WMCA to SiW-Mv2 silicone mask left out | ACER: 14.82 -> 4.79 | Improved |
| add diverse spoof prompt regularization l_div (ablation) | Ours · WMCA to SiW-Mv2 silicone mask left out | ACER: 10.86 -> 4.79 | Improved |
| add contrastive generation regularization l_con (ablation) | Ours · WMCA to SiW-Mv2 silicone mask left out | ACER: 6.51 -> 4.79 | Improved |
| add guidance regularization l_gui (ablation) | Ours · WMCA to SiW-Mv2 silicone mask left out | ACER: 7.47 -> 4.79 | Improved |
| add learnable unknown spoof prompts (ablation) | Ours · WMCA to SiW-Mv2 silicone mask left out | ACER: 9.14 -> 4.79 | Improved |
| add prior spoof prompts (ablation) | Ours · WMCA to SiW-Mv2 silicone mask left out | ACER: 5.24 -> 4.79 | Improved |
| use learned unknown spoof prompts from real faces only instead of CoOp trained with spoof images | CoOp · WMCA to SiW-Mv2 (avg of 14) | ACER: 30.01 -> 18.89 | Improved |
| use our method instead of OC-SCMNet | OC-SCMNet · WMCA to SiW-Mv2 (avg of 14) | ACER: 26.23 -> 18.89 | Improved |
| increase number of unknown spoof prompts from 4 to 20 (ablation) | Ours · WMCA to SiW-Mv2 silicone mask left out | all metrics | Mixed |
| use ours on obfuscation makeup | Ours vs prior methods · WMCA to SiW-Mv2 obfuscation makeup | ACER: 38.14 vs 33.06 | Hurt |
Closest papers by meaning
- CFPL-FAS: Class Free Prompt Learning for Generalizable Face Anti-spoofing (CVPR 2024 · similarity 0.91)
- FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models (AAAI 2026 · similarity 0.89)
- Interpretable Face Anti-Spoofing: Enhancing Generalization with Multimodal Large Language Models (AAAI 2025 · similarity 0.89)
- From Intuition to Investigation: A Tool-Augmented Reasoning MLLM Framework for Generalizable Face Anti-Spoofing (CVPR 2026 · similarity 0.89)
- SLIP: Spoof-Aware One-Class Face Anti-Spoofing with Language Image Pretraining (AAAI 2025 · similarity 0.89)
- Test-Time Domain Generalization for Face Anti-Spoofing (CVPR 2024 · 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.
