Research map / 3D vision / Face and shape
StyleSDF: High-Resolution 3D-Consistent Image and Geometry Generation
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 |
|---|---|---|---|
| Use StyleSDF instead of GIRAFFE | FFHQ image generation · FFHQ | FID: 31.2 -> 11.5 | Helped |
| Use StyleSDF instead of GIRAFFE | FFHQ image generation · FFHQ | KID (×1000): 20.1 -> 2.65 | Helped |
| Use StyleSDF instead of PiGAN | AFHQ image generation · AFHQ | FID: 52.4 -> 12.8 | Helped |
| Use StyleSDF instead of GIRAFFE | AFHQ image generation · AFHQ | KID (×1000): 15.1 -> 4.47 | Helped |
| Use SDF-based geometry instead of PiGAN's opacity field | StyleSDF volume rendering · FFHQ | modified Chamfer distance: 11.04 -> 0.40 | Helped |
| Use SDF-based geometry instead of PiGAN's opacity field | StyleSDF volume rendering · AFHQ | modified Chamfer distance: 8.66 -> 0.63 | Helped |
| Use SDF-based geometry instead of PiGAN's opacity field | StyleSDF volume renderer · FFHQ | RGB reprojection error: 14.7 -> 2.9 | Helped |
| Use SDF-based geometry instead of PiGAN's opacity field | StyleSDF volume renderer · AFHQ | RGB reprojection error: 16.5 -> 2.6 | Helped |
| Add the minimal surface loss (ablation) | StyleSDF volume renderer · FFHQ | geometry quality | Helped |
| Add sphere initialization (ablation) | StyleSDF volume renderer · FFHQ | geometry quality | Helped |
| Attach Gaussian noise to the 3D surface (ablation) | StyleSDF video generation · AFHQ | video flickering | Helped |
| Attach Gaussian noise to the 3D surface (ablation) | StyleSDF video generation · AFHQ | individual-frame image quality | Hurt |
Closest papers by meaning
- Dual Mapping of 2D StyleGAN for 3D-Aware Image Generation and Manipulation (Student Abstract) (AAAI 2024 · similarity 0.90)
- StyleNeRF: A Style-based 3D Aware Generator for High-resolution Image Synthesis (ICLR 2022 · similarity 0.90)
- What You See is What You GAN: Rendering Every Pixel for High-Fidelity Geometry in 3D GANs (CVPR 2024 · similarity 0.87)
- ViewDiff: 3D-Consistent Image Generation with Text-to-Image Models (CVPR 2024 · similarity 0.85)
- Mesh BDF: Barycentric Dominance Field for 3D Native Mesh Generation (NeurIPS 2026 · similarity 0.85)
- Structured 3D Latents for Scalable and Versatile 3D Generation (CVPR 2025 · 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.