Research map / Label-efficient and robust learning / Recognition and long-tail
Feature Generation and Hypothesis Verification for Reliable Face Anti-spoofing
Shice Liu, Shitao Lu, Hongyi Xu, Jing Yang, Shouhong Ding, Lizhuang Ma
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 FGHV rather than the strongest listed baseline | cross-dataset face anti-spoofing · O&C&I to M | HTER (%): 10.83 -> 9.17 | Improved |
| Use FGHV rather than the strongest listed baseline | cross-dataset face anti-spoofing · O&M&I to C | HTER (%): 15.21 -> 12.47 | Improved |
| Use FGHV rather than the strongest listed baseline | cross-dataset face anti-spoofing · O&C&M to I | HTER (%): 11.63 -> 16.29 | Hurt |
| Use FGHV rather than the strongest listed baseline | cross-dataset face anti-spoofing · I&C&M to O | HTER (%): 13.16 -> 13.58 | Hurt |
| Use both real-face and known-attack feature generators with all three constraints (ablation) | FGHV · O&C&I to M | HTER (%): 25.00 -> 9.17 | Improved |
| Add the real-face feature generator (ablation) | FGHV with only known-attack FGN · O&C&I to M | HTER (%): 20.00 -> 9.17 | Improved |
| Add the known-attack feature generator (ablation) | FGHV with only real-face FGN · O&C&I to M | HTER (%): 15.83 -> 9.17 | Improved |
| Add VAR to RCC (ablation) | FGHV · O&C&I to M | HTER (%): 14.16 -> 10.00 | Improved |
| Add DDC to RCC (ablation) | FGHV · O&C&I to M | HTER (%): 14.16 -> 10.83 | Improved |
| Add RCC to VAR+DDC (ablation) | FGHV · O&C&I to M | HTER (%): 12.50 -> 9.17 | Improved |
| Use VAR alone instead of RCC (ablation) | FGHV · O&C&I to M | HTER (%): 14.16 -> 33.33 | Hurt |
| Increase the number of sampled latent vectors (ablation) | FGHV · I&C&M to O | AUC | Improved |
Closest papers by meaning
- Test-Time Domain Generalization for 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 0.90)
- Gradient Alignment for Cross-Domain Face Anti-Spoofing (CVPR 2024 · similarity 0.90)
- FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models (AAAI 2026 · similarity 0.89)
- One-Class Face Anti-spoofing via Spoof Cue Map-Guided Feature Learning (CVPR 2024 · similarity 0.89)
- SLIP: Spoof-Aware One-Class Face Anti-Spoofing with Language Image Pretraining (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.
