Research map / 3D vision / Depth and stereo
RealX3D: A Physically-Degraded 3D Benchmark for Multi-view Visual Restoration and Reconstruction
Shuhong Liu, Chenyu Bao, Ziteng Cui, Yun Liu, Xuangeng Chu, Lin Gu, Marcos V. Conde, Ryo Umagami, Tomohiro Hashimoto, Zijian Hu, Tianhan Xu, Yuan Gan and 2 more
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 blur-aware DeBlurring-3DGS instead of vanilla 3DGS | 3DGS · RealX3D motion blur (NVS) | PSNR: 19.41 -> 20.32 | Improved |
| use blur-aware methods instead of vanilla 3DGS | 3DGS · RealX3D defocus blur (NVS) | PSNR: 19.79 -> 20.58 | Mixed |
| use DeBlurring-3DGS instead of vanilla 3DGS | 3DGS · RealX3D defocus blur (NVS) | PSNR: 19.79 -> 19.47 | Hurt |
| use SpotLessSplats instead of vanilla 3DGS | 3DGS · RealX3D dynamic occlusion (NVS) | PSNR: 19.83 -> 26.28 | Improved |
| use SpotLessSplats instead of vanilla 3DGS | 3DGS · RealX3D reflection (NVS) | PSNR: 21.60 -> 24.52 | Improved |
| use Wild3A instead of vanilla 3DGS | 3DGS · RealX3D dynamic occlusion (NVS) | PSNR: 19.83 -> 15.67 | Hurt |
| use LITA-GS instead of vanilla 3DGS | 3DGS · RealX3D low-light (NVS) | PSNR: 6.66 -> 15.57 | Improved |
| use Luminance-GS instead of vanilla 3DGS | 3DGS · RealX3D varying exposure (NVS) | PSNR: 7.39 -> 14.51 | Improved |
| use SeaThru-NeRF instead of vanilla 3DGS | 3DGS · RealX3D smoke (NVS) | PSNR: 9.76 -> 7.55 | Hurt |
| use Watersplatting instead of vanilla 3DGS | 3DGS · RealX3D smoke (NVS) | PSNR: 9.76 -> 10.78 | Improved |
| degrade input views with real degradations | VGGT · RealX3D | AUC@10: 92.98 -> 91.32 | Hurt |
| degrade input views with real degradations | DepthAnything3 · RealX3D | AUC@10: 94.87 -> 79.26 | Hurt |
Closest papers by meaning
- 3DReflecNet: A Large-Scale Dataset for 3D Reconstruction of Reflective, Transparent, and Low-Texture Objects (CVPR 2026 · similarity 0.89)
- Scene4U: Hierarchical Layered 3D Scene Reconstruction from Single Panoramic Image for Your Immerse Exploration (CVPR 2025 · similarity 0.88)
- UniK3D: Universal Camera Monocular 3D Estimation (CVPR 2025 · similarity 0.88)
- BLURD: Benchmarking and Learning using a Unified Rendering and Diffusion Model (NeurIPS 2024 · similarity 0.87)
- SPAR3D: Stable Point-Aware Reconstruction of 3D Objects from Single Images (CVPR 2025 · similarity 0.87)
- Scalable Benchmarking and Robust Learning for Noise-Free Ego-Motion and 3D Reconstruction from Noisy Video (ICLR 2025 · similarity 0.87)
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.
