Research map / 3D vision / 3D generation
MV2UV: Generating High-quality UV Texture Maps with Multiview Prompts
Zheng Zhang, Qinchuan Zhang, Yuteng Ye, Zhi Chen, Penglei Ji, Mengfei Li, Wenxiao Zhang, Yuan Liu
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 geometric position encoding to attention layers (ablation) | MV2UV · GSO | PSNR: 13.6 -> 25.7 | Improved |
| add geometric position encoding to attention layers (ablation) | MV2UV · DTC | PSNR: 13.9 -> 28.8 | Improved |
| add reference attention on multiview features (ablation) | MV2UV · GSO | PSNR: 18.9 -> 25.7 | Improved |
| add reference attention on multiview features (ablation) | MV2UV · DTC | PSNR: 22.6 -> 28.8 | Improved |
| add UV self attention (ablation) | MV2UV · GSO | PSNR: 21.5 -> 25.7 | Improved |
| add UV self attention (ablation) | MV2UV · DTC | SSIM: 0.863 -> 0.886 | Improved |
| replace the back view with an inconsistent MV-Adapter regeneration (strength 0.5) (ablation) | MV2UV · GSO | PSNR: 25.7 -> 25.5 | No clear effect |
| replace the back view with an inconsistent MV-Adapter regeneration (strength 0.5) (ablation) | MV2UV · DTC | SSIM: 0.886 -> 0.885 | No clear effect |
| use MV2UV UV-space generation instead of MV-Adapter back-projection on occluded regions | MV-Adapter · Occluded regions (10 samples) | FID: 123.6 -> 67.5 | Improved |
| use MV2UV UV-space generation instead of MV-Adapter back-projection on occluded regions | MV-Adapter · Occluded regions (10 samples) | KID: 157.9 -> 47.6 | Improved |
| use MV2UV instead of TEXGen | TEXGen · GSO | FID: 75.2 -> 24.4 | Improved |
| use MV2UV instead of MV-Adapter | MV-Adapter · GSO | FID: 24.7 -> 24.4 | No clear effect |
Closest papers by meaning
- Paint3D: Paint Anything 3D with Lighting-Less Texture Diffusion Models (CVPR 2024 · similarity 0.88)
- AUV-Net: Learning Aligned UV Maps for Texture Transfer and Synthesis (CVPR 2022 · similarity 0.88)
- TUVF: Learning Generalizable Texture UV Radiance Fields (ICLR 2024 · similarity 0.86)
- ViewDiff: 3D-Consistent Image Generation with Text-to-Image Models (CVPR 2024 · similarity 0.85)
- Kiss3DGen: Repurposing Image Diffusion Models for 3D Asset Generation (CVPR 2025 · similarity 0.85)
- Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cycle (AAAI 2025 · similarity 0.84)
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.
