Research map / 3D vision / Face and shape
3DStyleNet: Creating 3D Shapes With Geometric and Texture Style Variations
Kangxue Yin, Jun Gao, Maria Shugrina, Sameh Khamis, Sanja Fidler
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 3DStyleNet geometry and texture augmentation versus no augmentation | DISN · 3D reconstruction, animals, Unseen Shapes | Chamfer distance: 0.065 -> 0.037 | Improved |
| add random affine augmentation | DISN · 3D reconstruction, animals, Seen Shapes | Chamfer distance: 0.025 -> 0.040 | Hurt |
| add joint test-time fine-tuning of geometry and texture networks (ablation) | 3DStyleNet · 3D reconstruction, animals, Unseen Shapes | Chamfer distance: 0.039 -> 0.037 | Improved |
| add joint test-time fine-tuning of geometry and texture networks (ablation) | 3DStyleNet · 3D reconstruction, animals, Unseen Shapes | F-score: 0.183 -> 0.218 | Improved |
| add texture transfer to geometry transfer only (ablation) | 3DStyleNet · 3D reconstruction, animals, Unseen Shapes | Chamfer distance: 0.040 -> 0.037 | Improved |
| add geometry transfer to texture transfer only (ablation) | 3DStyleNet · 3D reconstruction, animals, Unseen Shapes | Chamfer distance: 0.051 -> 0.037 | Improved |
| add geometry transfer to texture transfer only (ablation) | 3DStyleNet · 3D reconstruction, animals, Seen Shapes | Chamfer distance: 0.012 -> 0.016 | Hurt |
| use part-aware affine geometry transfer instead of Neural Cage geometry | DISN · 3D reconstruction, animals, Unseen Shapes | Chamfer distance: 0.044 -> 0.040 | Improved |
| use 3DStyleNet full instead of random COCO texture augmentation | DISN · 3D reconstruction, animals, Unseen Shapes | Chamfer distance: 0.045 -> 0.037 | Improved |
| use 3DStyleNet full instead of Neural Cage + Style Transfer | DISN · 3D reconstruction, animals, Unseen Shapes | F-score: 0.181 -> 0.218 | Improved |
| use 3DStyleNet versus Neural Cage + Style Transfer on Chamfer | DISN · 3D reconstruction, animals, Unseen Shapes | Chamfer distance: 0.040 -> 0.037 | Improved |
| use 3DStyleNet instead of Neural Cage plus style transfer in user study | 3DStyleNet · AMT user study | users closer to target: 41.9% vs 26.8% | Improved |
Closest papers by meaning
- GET3D: A Generative Model of High Quality 3D Textured Shapes Learned from Images (NeurIPS 2022 · similarity 0.90)
- 3DToonify: Creating Your High-Fidelity 3D Stylized Avatar Easily from 2D Portrait Images (CVPR 2024 · similarity 0.88)
- Styl3R: Instant 3D Stylized Reconstruction for Arbitrary Scenes and Styles (NeurIPS 2025 · similarity 0.88)
- LeGO: Leveraging a Surface Deformation Network for Animatable Stylized Face Generation with One Example (CVPR 2024 · similarity 0.88)
- TextureDreamer: Image-Guided Texture Synthesis Through Geometry-Aware Diffusion (CVPR 2024 · similarity 0.87)
- Wukong's 72 Transformations: High-fidelity Textured 3D Morphing via Flow Models (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.
