Research map / Image generation / Editing and style transfer
Improving Fairness in Facial Albedo Estimation via Visual-Textual Cues
Read the paperCheck 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 |
|---|---|---|---|
| Use ID2Albedo instead of TRUST (BalancedAlb) | Facial albedo estimation · FAIR benchmark | Avg. ITA: 13.87 -> 12.07 | Helped |
| Use ID2Albedo instead of TRUST (BalancedAlb) | Facial albedo estimation · FAIR benchmark | Bias: 2.79 -> 4.91 | Hurt |
| Use ID2Albedo instead of TRUST | Facial albedo estimation · FFHQ | M-SSIM: 0.64 -> 0.87 | Helped |
| Use a subspace-based GAN rather than StyleGANv2-ADA (ablation) | Albedo generator · Aligned UV albedo data | FID: 49.3 -> 42.2 | Helped |
| Freeze the ArcFace encoder rather than fine-tuning it (ablation) | ID2Albedo · FAIR benchmark | Avg. ITA: 14.58 -> 13.46 | Helped |
| Use illumination-robust ArcFace identity features instead of a scratch-trained ResNet-100 (ablation) | Albedo encoder · FAIR benchmark | Avg. ITA: 58.46 -> 13.46 | Helped |
| Add CLIP-based multiple-attribute cues (ablation) | ID2Albedo · FAIR benchmark | Avg. ITA: 25.66 -> 13.46 | Helped |
| Add CLIP cues for attributes beyond race (ablation) | ID2Albedo · FAIR benchmark | Avg. ITA: 16.21 -> 13.46 | Helped |
| Replace CLIP cues with manually labeled race cues (ablation) | ID2Albedo · FAIR benchmark | Avg. ITA: 13.46 -> 18.13 | Hurt |
Closest papers by meaning
- A Morphable Face Albedo Model (CVPR 2020 · similarity 0.90)
- S^3-Face: SSS-Compliant Facial Reflectance Estimation via Diffusion Priors (CVPR 2025 · similarity 0.88)
- Monocular Identity-Conditioned Facial Reflectance Reconstruction (CVPR 2024 · similarity 0.88)
- WildCap: Facial Albedo Capture in the Wild via Hybrid Inverse Rendering (CVPR 2026 · similarity 0.88)
- SFDM: Robust Decomposition of Geometry and Reflectance for Realistic Face Rendering from Sparse-view Images (CVPR 2025 · similarity 0.84)
- 3D Facial Expressions through Analysis-by-Neural-Synthesis (CVPR 2024 · similarity 0.84)
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.