Research map / Image generation / Text-to-image diffusion
MatPedia: A Universal Generative Foundation for High-Fidelity Material Synthesis
Di Luo, Shuhui Yang, Mingxin Yang, Jiawei Lu, Yixuan Tang, Xintong Han, Zhuo Chen, Beibei Wang, Chunchao Guo
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 RGB-only images (mixed training) (ablation) | MatPedia text-to-material · text-to-material | CLIP score: 0.275 -> 0.283 | Improved |
| add RGB-only images (mixed training) (ablation) | MatPedia text-to-material · text-to-material | DINO-FID: 1.62 -> 1.31 | Improved |
| fine-tune the 3D VAE decoder on PBR data (ablation) | MatPedia VAE · PBR reconstruction | PSNR Normal: 27.29 -> 30.84 | Improved |
| fine-tune the 3D VAE decoder on PBR data (ablation) | MatPedia VAE · PBR reconstruction | PSNR Roughness: 31.36 -> 36.56 | Improved |
| fine-tune the 3D VAE decoder on PBR data (ablation) | MatPedia VAE · PBR reconstruction | PSNR Render: 33.12 -> 34.25 | Improved |
| fine-tune the 3D VAE decoder on PBR data (ablation) | MatPedia VAE · PBR reconstruction | PSNR Basecolor: 29.27 -> 31.30 | Improved |
| fine-tune the 3D VAE decoder on PBR data (ablation) | MatPedia VAE · PBR reconstruction | PSNR Metallic: 42.45 -> 45.59 | Improved |
| use MatPedia instead of MatFuse | text-to-material · text-to-material | DINO-FID: 1.90 -> 1.31 | Improved |
| use MatPedia instead of MatFuse | image-to-material basecolor · image-to-material | DINO score: 0.649 -> 0.907 | Improved |
| use MatPedia instead of Material Palette | material decomposition basecolor · material decomposition | MSE: 0.058 -> 0.009 | Improved |
| use MatPedia instead of RGB<->X | material decomposition roughness · material decomposition | LPIPS: 0.622 -> 0.423 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MatFuse — MatFuse: Controllable Material Generation with Diffusion Models (CVPR 2024) · better on 10 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Material Palette: Extraction of Materials from a Single Image (CVPR 2024 · similarity 0.92)
- MatFuse: Controllable Material Generation with Diffusion Models (CVPR 2024 · similarity 0.89)
- IntrinsiX: High-Quality PBR Generation using Image Priors (NeurIPS 2025 · similarity 0.89)
- TexGaussian: Generating High-quality PBR Material via Octree-based 3D Gaussian Splatting (CVPR 2025 · similarity 0.89)
- MatSynth: A Modern PBR Materials Dataset (CVPR 2024 · similarity 0.89)
- MeshGen: Generating PBR Textured Mesh with Render-Enhanced Auto-Encoder and Generative Data Augmentation (CVPR 2025 · 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.
