Research map / 3D vision / Face and shape
Differential 3D Facial Recognition: Adding 3D to Your State-of-the-Art 2D Method
J. Matías Di Martino, Fernando Suzacq, Mauricio Delbracio, Qiang Qiu, Guillermo Sapiro
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 3D depth features to RGB features (gamma=0.3) | RGB baseline · ND-2006 | Rank-1 accuracy: 78.5 -> 90.6 | Improved |
| weight depth more (gamma 0.3 to 0.8) (ablation) | 2D+3D features · ND-2006 | Rank-1 accuracy: 90.6 -> 85.0 | Hurt |
| add 3D depth features to RGB features | RGB baseline · ND-2006 spoofing | TPR at FPR=1e-3: 21.8 -> 85.5 | Improved |
| add RGB features to depth features (gamma=0.3, beta=0.35, alpha=5) | Depth baseline · ND-2006 spoofing | TPR at FPR=1e-3: 88.4 -> 83.8 | Hurt |
| add 3D features to 2D features under strong ambient light (power=200%) | RGB baseline · ND-2006 ambient light | Rank-5 accuracy: 53.9 -> 80.7 | Improved |
| add 3D features to ArcFace | ArcFace · ND-2006 spoofing | TPR at FPR=1e-3: 0 -> 84.7 | Improved |
| add 3D features to ArcFace | ArcFace · ND-2006 | Rank-1 accuracy: 100 -> 100 | No clear effect |
| raise beta threshold from 0.35 to 0.5 in non-linear distance (alpha=10) (ablation) | 2D+3D features · ND-2006 spoofing | ACER: 3.9 -> 5.9 | Hurt |
Closest papers by meaning
- Physical-World Optical Adversarial Attacks on 3D Face Recognition (CVPR 2023 · similarity 0.84)
- Towards Effective Adversarial Textured 3D Meshes on Physical Face Recognition (CVPR 2023 · similarity 0.83)
- Enhancing Face Recognition With Self-Supervised 3D Reconstruction (CVPR 2022 · similarity 0.83)
- StrokeFaceNeRF: Stroke-based Facial Appearance Editing in Neural Radiance Field (CVPR 2024 · similarity 0.81)
- ProjAttacker: A Configurable Physical Adversarial Attack for Face Recognition via Projector (CVPR 2025 · similarity 0.81)
- Learning to Decouple the Lights for 3D Face Texture Modeling (NeurIPS 2024 · similarity 0.81)
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.
