Research map / Label-efficient and robust learning / Recognition and long-tail
Spherical Confidence Learning for Face Recognition
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 SCF confidence learning to CosFace with ResNet34 | CosFace · IJB-B | TAR@FAR=1e-5: 84.14 -> 86.55 | Helped |
| Add SCF confidence learning to CosFace with ResNet34 | CosFace · IJB-C | TAR@FAR=1e-5: 88.72 -> 90.82 | Helped |
| Add SCF confidence learning to CosFace with ResNet100 | CosFace · IJB-B | TAR@FAR=1e-4: 94.59 -> 94.95 | Helped |
| Add SCF confidence learning to CosFace with ResNet100 | CosFace · IJB-C | TAR@FAR=1e-4: 95.95 -> 96.22 | Helped |
| Add SCF confidence learning to CosFace with ResNet34 | CosFace · CFP-FP | accuracy: 94.17 -> 95.19 | Helped |
| Add SCF confidence learning to CosFace with ResNet34 | CosFace · CPLFW | accuracy: 89.50 -> 90.68 | Helped |
| Add SCF confidence learning to CosFace with ResNet100 | CosFace · CPLFW | accuracy: 92.75 -> 93.26 | Helped |
| Replace spherical r-vMF confidence modeling with independent Gaussian modeling (SCF-G) (ablation) | SCF · IJB-B | TAR@FAR=1e-5: 86.55 -> 84.45 | Hurt |
| Replace spherical r-vMF confidence modeling with independent Gaussian modeling (SCF-G) (ablation) | SCF · IJB-C | TAR@FAR=1e-5: 90.82 -> 89.04 | Hurt |
| Use PFE's pairwise mutual-likelihood objective with spherical r-vMF (ablation) | Spherical confidence learning · Training | training stability | Hurt |
| Use independent Gaussian uncertainty instead of spherical confidence modeling (ablation) | PFE · IJB-B | TAR@FAR=1e-5: 86.55 vs 84.36 | Hurt |
| Use independent Gaussian uncertainty instead of spherical confidence modeling (ablation) | PFE · IJB-C | TAR@FAR=1e-5: 90.82 vs 89.13 | Hurt |
Closest papers by meaning
- Data Uncertainty Learning in Face Recognition (CVPR 2020 · similarity 0.86)
- SphereFace: Deep Hypersphere Embedding for Face Recognition (CVPR 2017 · similarity 0.86)
- Attributes Shape the Embedding Space of Face Recognition Models (ICML 2025 · similarity 0.85)
- SPHERE-JEPA: Spherical Prediction with Homogeneous Embeddings (NeurIPS 2026 · similarity 0.83)
- HyperFace: Generating Synthetic Face Recognition Datasets by Exploring Face Embedding Hypersphere (ICLR 2025 · similarity 0.83)
- LAFS: Landmark-based Facial Self-supervised Learning for Face Recognition (CVPR 2024 · similarity 0.82)
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.