Huawei's SUPRA-Tex turns one image into relightable 3D materials with WYSIWYG fidelity
Chen Yang · hf · 2026-10-16
SUPRA-Tex, from Huawei, SJTU and HUST, tackles single-image 3D material generation by decoupling visible-material estimation, surface alignment, and observation-guided 3D completion. An MLRM-based image editing model predicts lighting-separated PBR maps in the reference view; these are projected onto visible surfaces and encoded via a pretrained material VAE into observation latents. An interleaved flow transformer holds them as read-only anchors while alternating unobserved-region and full-surface updates, achieving WYSIWYG fidelity on visible regions and coherent completion elsewhere. Trained only on public data, it reaches SOTA performance.
More from Multimodal
- Tripo hosts first world-building hackathon in SF with World Labs and Runway talks — mhdfaran · 2026-10-16
- Kling 4.0 boosts text-to-video with character consistency and lip sync — CurieuxExplorer · 2026-10-16
- 5.8M-param Speech Event Detector adds laughs, breaths and背景 sounds to Whisper transcripts — Ayush Kumar Bar · 2026-10-16
- Giant Michael Jackson AI fashion images show the power of extreme proportions — OVolosin82152 · 2026-10-16
- BFL says Flux 3 Klein complaints stem from a Comfy workflow bug, fix underway — desktop4070 · 2026-10-16
- MiniMax H3 + RVC-project combo delivers surprisingly good voice results — ColdComplaint8 · 2026-10-16