Research map / Image generation / Text-to-image diffusion
MaTe: Images Are All You Need for Material Transfer via Diffusion Transformer
Nisha Huang, Henglin Liu, Yizhou Lin, Kaer Huang, Chubin Chen, Jie Guo, Tong-yee Lee, Xiu Li
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 |
|---|---|---|---|
| MaTe versus best baseline MaterialFusion | MaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects) | SSIM: 0.8263 -> 0.8617 | Improved |
| MaTe versus best baseline MaterialFusion | MaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects) | LPIPS: 0.1565 -> 0.1319 | Improved |
| MaTe versus best baseline ZeST | MaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects) | CLIP similarity: 0.8627 -> 0.8825 | Improved |
| MaTe versus MaterialFusion in user study overall | MaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects) | user overall preference: 10.48% -> 83.82% | Improved |
| MaTe versus MaterialFusion in inference time | MaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects) | inference time (s): 43 -> 11 | Improved |
| replace ControlNet (1.59B params) with FLUX depth LoRA (620M) (ablation) | MaTe · material transfer | inference speedup: 1.8x speedup | Improved |
| increase Cross-Bias strength gamma toward 1.8 (ablation) | MaTe · qualitative | material transfer effect: 0 -> 1.8 | Improved |
| increase gamma beyond 1.8 (ablation) | MaTe · qualitative | material transfer effect | No clear effect |
| vary depth LoRA weight w in 0.7 to 1.0 (ablation) | MaTe · qualitative | depth influence: 0.7 to 1.0 | Improved |
| use illumination image instead of original input image or random noise (ablation) | MaTe · qualitative | lighting and color fidelity | Improved |
| raise CFG strength from 10 to 30 (ablation) | MaTe · qualitative | detail and fidelity: 10 -> 30 | Improved |
| raise CFG strength beyond 30 (up to 50) (ablation) | MaTe · qualitative | image quality: 30 -> 50 | Hurt |
Closest papers by meaning
- Material Anything: Generating Materials for Any 3D Object via Diffusion (CVPR 2025 · similarity 0.86)
- FAM Diffusion: Frequency and Attention Modulation for High-Resolution Image Generation with Stable Diffusion (CVPR 2025 · similarity 0.86)
- Plug-and-Play Diffusion Distillation (CVPR 2024 · similarity 0.86)
- Material Palette: Extraction of Materials from a Single Image (CVPR 2024 · similarity 0.86)
- Simpler Diffusion: 1.5 FID on ImageNet512 with Pixel-space Diffusion (CVPR 2025 · similarity 0.86)
- Diffusion Renderer: Neural Inverse and Forward Rendering with Video Diffusion Models (CVPR 2025 · similarity 0.86)
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.
