Research map / 3D vision / 3D generation
Auto3R: Automated 3D Reconstruction and Scanning via Data-driven Uncertainty Quantification
Chentao Shen, Sizhe Zheng, Bingqian Wu, Yaohua Feng, Yuanchen Fei, Mingyu Mei, Hanwen Jiang, Xiangru Huang
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 |
|---|---|---|---|
| use Auto3R UQ instead of AVS | active view selection · Objaverse objects | PSNR avg: 23.94 -> 27.00 | Improved |
| use Auto3R UQ instead of AVS | active view selection · Mip-NeRF360 | SSIM avg: 0.6062 vs 0.6055 | No clear effect |
| use Auto3R UQ instead of AVS | active view selection · Mip-NeRF360 | PSNR worst-5%: 9.28 -> 10.75 | Improved |
| add depth-aware blending (ablation) | Auto3R UQ · Objaverse | PSNR avg: 22.60 -> 23.41 (table); text states 22.48 -> 23.41 | Mixed |
| add depth-uncertainty reweighting (ablation) | Auto3R UQ · Objaverse | PSNR avg: 23.41 -> 23.77 | Improved |
| quadratic instead of linear depth weighting (ablation) | Auto3R UQ · Objaverse | PSNR avg: 23.77 -> 22.59 | Hurt |
| combine Auto3R with GIR inverse-rendering framework | Auto3R · specular objects | PSNR avg: 26.41 -> 30.29 | Improved |
| use sequence-based path UQ extension | Auto3R · real-robot scanning | PSNR: 22.91 -> 28.56 | Improved |
| use analytic 3D uncertainty (GAUSS-MI) instead of learned UQ | active view selection · Objaverse | PSNR avg: 27.00 -> 16.25 | Hurt |
| adapt generic image-quality metrics (MANIQA/MUSIQ) for view selection | active view selection · Objaverse | PSNR avg: 16.09 (Random) vs 14.34-15.82 | No clear effect |
Closest papers by meaning
- Variational Multi-scale Representation for Estimating Uncertainty in 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.85)
- SPAR3D: Stable Point-Aware Reconstruction of 3D Objects from Single Images (CVPR 2025 · similarity 0.85)
- BAT3R: Robust Online 3D Reconstruction with Bayesian Adaptive State Updates (NeurIPS 2026 · similarity 0.84)
- DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024 · similarity 0.84)
- Fin3R: Fine-tuning Feed-forward 3D Reconstruction Models via Monocular Knowledge Distillation (NeurIPS 2025 · similarity 0.84)
- Test3R: Learning to Reconstruct 3D at Test Time (NeurIPS 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.
