Research map / 3D vision / 3D generation
Decorate3D: Text-Driven High-Quality Texture Generation for Mesh Decoration in the Wild
Yanhui Guo, Xinxin Zuo, Peng Dai, Juwei Lu, Xiaolin Wu, Li Cheng, Youliang Yan, Songcen Xu, Xiaofei Wu
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 |
|---|---|---|---|
| initialize UV neural texture with MLPD prior instead of random initialization (ablation) | Decorate3D · qualitative texture generation | convergence speed and geometry match | Improved |
| add depth guidance (structure-aware SDS) over naive SDS (ablation) | Decorate3D · qualitative texture generation | geometry-matching texture | Improved |
| add few-view resampling (FVR) training instead of per-view generation (ablation) | Decorate3D · multi-view consistency | 3D consistency | Improved |
| feed neural texture into diffusion VAE instead of FVR-trained MLP (ablation) | Decorate3D · qualitative texture generation | texture sharpness | Hurt |
| add super-resolution on UV texture (ablation) | Decorate3D · qualitative texture generation | texture clarity | Improved |
| increase FVR hyperparameter N to a large value (ablation) | Decorate3D · qualitative texture generation | texture sharpness | Hurt |
| add neural renderer to convert neural UV texture to RGB (ablation) | Decorate3D · qualitative texture generation | color quality | Improved |
| Decorate3D versus TEXTure on NIQE | Decorate3D · Table 1 NIQE | NIQE: 14.83 -> 14.82 | Improved |
| Decorate3D versus CLIP-Mesh on CLIP B/32 R-Precision | Decorate3D · Table 1 CLIP B/32 | R-Precision: 31.95 -> 30.42 | Hurt |
| Decorate3D versus DreamFusion on CLIP L/14 R-Precision | Decorate3D · Table 1 CLIP L/14 | R-Precision: 24.38 -> 24.89 | Improved |
| Decorate3D versus Latent-Paint on CLIP B/16 R-Precision | Decorate3D · Table 1 CLIP B/16 | R-Precision: 26.19 -> 30.47 | Improved |
Closest papers by meaning
- Paint3D: Paint Anything 3D with Lighting-Less Texture Diffusion Models (CVPR 2024 · similarity 0.89)
- Magic123: One Image to High-Quality 3D Object Generation Using Both 2D and 3D Diffusion Priors (ICLR 2024 · similarity 0.89)
- PaintHuman: Towards High-Fidelity Text-to-3D Human Texturing via Denoised Score Distillation (AAAI 2024 · similarity 0.88)
- MeshGen: Generating PBR Textured Mesh with Render-Enhanced Auto-Encoder and Generative Data Augmentation (CVPR 2025 · similarity 0.88)
- SF3D: Stable Fast 3D Mesh Reconstruction with UV-unwrapping and Illumination Disentanglement (CVPR 2025 · similarity 0.88)
- One-2-3-45++: Fast Single Image to 3D Objects with Consistent Multi-View Generation and 3D Diffusion (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.
