Research map / 3D vision / Face and shape
A Shading-Guided Generative Implicit Model for Shape-Accurate 3D-Aware Image Synthesis
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 |
|---|---|---|---|
| replace global shading by local points-specific shading (ablation) | ShadeGAN · BFM | FID: 17.7 -> 30.1 | Hurt |
| replace global shading by local points-specific shading (ablation) | ShadeGAN · BFM | SIDE: 0.607 -> 0.754 | Hurt |
| replace global shading by local points-specific shading (ablation) | ShadeGAN · BFM | MAD: 14.52 -> 18.18 | Hurt |
| remove dependence of albedo on lighting (ablation) | ShadeGAN · BFM | FID: 17.7 -> 19.2 | Hurt |
| remove dependence of albedo on lighting (ablation) | ShadeGAN · BFM | MAD: 14.52 -> 14.53 | No clear effect |
| remove albedo dependence on viewpoint (ablation) | ShadeGAN · BFM | FID: 17.7 -> 18.6 | Hurt |
| remove albedo dependence on viewpoint (ablation) | ShadeGAN · BFM | MAD: 14.52 -> 14.88 | Hurt |
| use manually tuned lighting prior instead of data-driven prior (ablation) | ShadeGAN · BFM | FID: 17.7 -> 20.2 | Hurt |
| use manually tuned lighting prior instead of data-driven prior (ablation) | ShadeGAN · BFM | MAD: 14.52 -> 15.38 | Hurt |
| add efficient volume rendering via surface tracking (ablation) | ShadeGAN · BFM | SIDE: 0.607 -> 0.673 | Hurt |
| add efficient volume rendering via surface tracking (ablation) | ShadeGAN · CelebA | training time (h): 92.3 -> 70.2 | Helped |
| add efficient volume rendering via surface tracking (ablation) | ShadeGAN · CelebA | inference time (s): 0.343 -> 0.179 | Helped |
Closest papers by meaning
- GRAF: Generative Radiance Fields for 3D-Aware Image Synthesis (NeurIPS 2020 · similarity 0.90)
- Generative Multiview Relighting for 3D Reconstruction under Extreme Illumination Variation (CVPR 2025 · similarity 0.88)
- GaussianCube: A Structured and Explicit Radiance Representation for 3D Generative Modeling (NeurIPS 2024 · similarity 0.87)
- Multistable Shape from Shading Emerges from Patch Diffusion (NeurIPS 2024 · similarity 0.87)
- RNG: Relightable Neural Gaussians (CVPR 2025 · similarity 0.87)
- Cue3D: Quantifying the Role of Image Cues in Single-Image 3D Generation (NeurIPS 2025 · similarity 0.87)
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.