SUPRA-Tex first trains an MLLM-based image editing model to predict lighting-separated PBR (physically based rendering) maps in the reference view in a single forward pass.
Grounding 3D texture generation in surface-aligned material observations improves fidelity
SUPRA-Tex decouples visible-material estimation from 3D completion, using projected PBR maps as fixed latent anchors to preserve reference details while coherently filling unobserved surfaces.
Chinese Tech
Chen Yang · Jiemin Fang · Fanpeng Meng · Zihao Huang · Jichen Pang · Guanjun Wu · +2 more
Huawei Inc. · Shanghai Jiao Tong University · Huazhong University of Science and Technology
Research Digest··2 min read
The authors introduce SUPRA-Tex, a framework for single-image 3D generation that separates material estimation, surface alignment, and geometry-conditioned completion.
Why this paper
From Huawei Inc. and 2 others
In one line
SUPRA-Tex generates 3D assets with super-aligned textures, preserving reference details and completing unseen surfaces.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§