Research map / 3D vision / Face and shape
On Disentangling Spoof Trace for Generic Face Anti-Spoofing
Yaojie Liu, Joel Stehouwer, Xiaoming Liu
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 GAN training with single-layer spoof trace (ablation) | ESR baseline · Oulu-NPU Protocol 1 | ACER (%): 2.6 -> 2.1 | Improved |
| use disentangled 4-element trace (D-GAN) instead of single-layer trace (ablation) | ESR+GAN · Oulu-NPU Protocol 1 | ACER (%): 2.1 -> 1.6 | Improved |
| add pixel loss on synthesized spoof (poor-quality traces) (ablation) | ESR+GAN · Oulu-NPU Protocol 1 | ACER (%): 2.1 -> 4.5 | Hurt |
| add pixel loss on synthesized spoof (good-quality traces) (ablation) | ESR+D-GAN · Oulu-NPU Protocol 1 | ACER (%): 1.6 -> 1.1 | Improved |
| classify spoof medium from estimated spoof traces instead of original images | AlexNet · Oulu-NPU Protocol 1 spoof medium classification (3-class) | accuracy (%): 86.3 -> 92.0 | Improved |
| classify spoof medium from estimated spoof traces instead of original images | AlexNet · Oulu-NPU Protocol 1 spoof medium classification (5-class) | accuracy (%): 80.6 -> 83.3 | Improved |
| use STDN instead of DTL for unknown spoof detection | STDN · SiW-M Protocol II | EER (%): 16.1 -> 12.0 | Improved |
| use STDN instead of DTL for unknown spoof detection | STDN · SiW-M Protocol II | ACER (%): 16.8 -> 14.2 | Improved |
| use STDN instead of DeSpoof | STDN · Oulu-NPU Protocol 4 | ACER (%): 5.6 -> 3.8 | Improved |
| use STDN instead of DeSpoof | STDN · Oulu-NPU Protocol 1 | ACER (%): 1.5 -> 1.1 | Improved |
| use STDN instead of Auxiliary on known spoof types | STDN · SiW-M Protocol I (known spoof, overall) | ACER (%): 6.3 -> 4.1 | Improved |
| use STDN instead of Meta-FAS-DR in cross-spoof protocol | STDN · SiW Protocol 3 | ACER (%): 7.7 -> 7.9 | No clear effect |
Closest papers by meaning
- Spoof Trace Disentanglement for Generic Face Anti-Spoofing (TPAMI 2022 · similarity 0.96)
- Learning Polysemantic Spoof Trace: A Multi-Modal Disentanglement Network for Face Anti-spoofing (AAAI 2023 · similarity 0.92)
- Face De-Spoofing: Anti-Spoofing via Noise Modeling (ECCV 2018 · similarity 0.90)
- One-Class Face Anti-spoofing via Spoof Cue Map-Guided Feature Learning (CVPR 2024 · similarity 0.88)
- SLIP: Spoof-Aware One-Class Face Anti-Spoofing with Language Image Pretraining (AAAI 2025 · similarity 0.87)
- 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.
