Research map / 3D vision / Face and shape
Geometry-Aware Single-Image Full-Body Human Relighting
Chaonan Ji, Tao Yu, Kaiwen Guo, Jingxin Liu, Yebin 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 ambient occlusion map input and refinement network to coarse ray-traced shading (ablation) | Full-body shading refinement · Shading (synthetic) | PSNR: 29.84 -> 32.86 | Improved |
| add ambient occlusion map input to the refinement network (vs none) (ablation) | Full-body refinement net · Shading (synthetic) | PSNR: 30.89 -> 32.86 | Improved |
| use ambient occlusion map instead of normal map as auxiliary input (ablation) | Full-body refinement net · Shading (synthetic) | MSE: 0.886 -> 0.729 | Improved |
| use proposed de-lighting network instead of vanilla UNet (ablation) | De-lighting network · Albedo (synthetic) | PSNR: 26.73 -> 30.27 | Improved |
| use proposed network instead of vanilla HRNet (ablation) | De-lighting network · Albedo (synthetic) | PSNR: 27.21 -> 30.27 | Improved |
| add skip connections to vanilla HRNet (ablation) | Vanilla HRNet · Albedo (synthetic) | PSNR: 27.21 -> 28.53 | Improved |
| add skip connections to vanilla HRNet (ablation) | Vanilla HRNet · Albedo (synthetic) | SSIM: 0.9550 -> 0.9767 | Improved |
| use full-stage de-lighting network instead of 5-stage version (ablation) | De-lighting network · Albedo (synthetic) | MSE: 1.728 -> 1.468 | Improved |
| add face refinement module using 3DMM geometry prior (ablation) | Face shading · Face relighting (qualitative) | visual quality | Improved |
| proposed method vs re-trained RH baseline | Full framework · Relighting (synthetic) | PSNR: 26.45 -> 30.20 | Improved |
| proposed method vs SFHR | Full framework · Relighting (synthetic) | MSE: 5.833 -> 1.311 | Improved |
Closest papers by meaning
- Learning Visibility Field for Detailed 3D Human Reconstruction and Relighting (CVPR 2023 · similarity 0.90)
- Comprehensive Relighting: Generalizable and Consistent Monocular Human Relighting and Harmonization (CVPR 2025 · similarity 0.89)
- UniRelight: Learning Joint Decomposition and Synthesis for Video Relighting (NeurIPS 2025 · similarity 0.89)
- Generative Multiview Relighting for 3D Reconstruction under Extreme Illumination Variation (CVPR 2025 · similarity 0.89)
- Relightable and Animatable Neural Avatars from Videos (AAAI 2024 · similarity 0.88)
- IntrinsicAvatar: Physically Based Inverse Rendering of Dynamic Humans from Monocular Videos via Explicit Ray Tracing (CVPR 2024 · similarity 0.88)
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.
