AI Paper Atlas

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

ICCV 2025 · Image generation › Text-to-image diffusion · approach: Compositional text-to-image guidance · 5 citations (OpenAlex)

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 changedWhereMeasuredEffect
MaTe versus best baseline MaterialFusionMaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects)SSIM: 0.8263 -> 0.8617Improved
MaTe versus best baseline MaterialFusionMaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects)LPIPS: 0.1565 -> 0.1319Improved
MaTe versus best baseline ZeSTMaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects)CLIP similarity: 0.8627 -> 0.8825Improved
MaTe versus MaterialFusion in user study overallMaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects)user overall preference: 10.48% -> 83.82%Improved
MaTe versus MaterialFusion in inference timeMaTe · Material Transfer Benchmark (MTB, 60 materials x 30 objects)inference time (s): 43 -> 11Improved
replace ControlNet (1.59B params) with FLUX depth LoRA (620M) (ablation)MaTe · material transferinference speedup: 1.8x speedupImproved
increase Cross-Bias strength gamma toward 1.8 (ablation)MaTe · qualitativematerial transfer effect: 0 -> 1.8Improved
increase gamma beyond 1.8 (ablation)MaTe · qualitativematerial transfer effectNo clear effect
vary depth LoRA weight w in 0.7 to 1.0 (ablation)MaTe · qualitativedepth influence: 0.7 to 1.0Improved
use illumination image instead of original input image or random noise (ablation)MaTe · qualitativelighting and color fidelityImproved
raise CFG strength from 10 to 30 (ablation)MaTe · qualitativedetail and fidelity: 10 -> 30Improved
raise CFG strength beyond 30 (up to 50) (ablation)MaTe · qualitativeimage quality: 30 -> 50Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy