Research map / 3D vision / Face and shape
HUST: High-Fidelity Unbiased Skin Tone Estimation via Texture Quantization
Zimin Ran, Xingyu Ren, Xiang An, Kaicheng Yang, Ziyong Feng, Jing Yang, Rolandos Alexandros Potamias, Linchao Zhu, Jiankang Deng
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 age-based grouping to the baseline (ablation) | HUST · CelebAMask-HQ | SSIM: 0.8627 -> 0.8917 | Improved |
| add group identity loss on top of age grouping (ablation) | HUST · CelebAMask-HQ | PSNR: 25.91 -> 29.14 | Improved |
| add group identity loss to the baseline (ablation) | HUST · CelebAMask-HQ | ID similarity: 0.5094 -> 0.7386 | Improved |
| add age grouping to the baseline (ablation) | HUST · CelebAMask-HQ | ID similarity: 0.5094 -> 0.5136 | No clear effect |
| combine age grouping and group identity loss (ablation) | HUST · CelebAMask-HQ | LPIPS: 0.1456 -> 0.1109 | Improved |
| use dual discriminator instead of latent-only (ablation) | UV texture encoder · Multi-PIE | PSNR: 25.47 -> 29.47 | Improved |
| use dual discriminator instead of image-only (ablation) | UV texture encoder · Multi-PIE | PSNR: 21.52 -> 29.47 | Improved |
| use HUST codebook trained on 0.5M high-res faces instead of FFHQ VQGAN (ablation) | VQGAN · UV Texture (Ostec) | PSNR: 28.39 -> 31.88 | Improved |
| use HUST codebook instead of FFHQ VQGAN (ablation) | VQGAN · Color Image (CelebAMask-HQ) | LPIPS: 0.2957 -> 0.2451 | Improved |
| use HUST codebook instead of FFHQ VQGAN (ablation) | VQGAN · UV Albedo (3D Scan Store) | PSNR: 29.49 -> 31.00 | Improved |
| use HUST instead of TRUST for albedo estimation | TRUST · FAIR benchmark | bias score: 2.79 -> 1.58 | Improved |
| use HUST instead of ID2Albedo, without albedo ground truth | ID2Albedo · FAIR benchmark | avg ITA error: 12.07 -> 11.20 | Improved |
Closest papers by meaning
- Improving Fairness in Facial Albedo Estimation via Visual-Textual Cues (CVPR 2023 · similarity 0.91)
- S^3-Face: SSS-Compliant Facial Reflectance Estimation via Diffusion Priors (CVPR 2025 · similarity 0.90)
- A Morphable Face Albedo Model (CVPR 2020 · similarity 0.89)
- Monocular Identity-Conditioned Facial Reflectance Reconstruction (CVPR 2024 · similarity 0.86)
- FreeUV: Ground-Truth-Free Realistic Facial UV Texture Recovery via Cross-Assembly Inference Strategy (CVPR 2025 · similarity 0.86)
- UV-IDM: Identity-Conditioned Latent Diffusion Model for Face UV-Texture Generation (CVPR 2024 · similarity 0.85)
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.
