Graph-Guided Scene Reconstruction from Images with 3D Gaussian Splatting
Chong Cheng Gaochao Song Yiyang Yao Qinzheng Zhou, Gangjian Zhang Hao Wang · 2025
Read Graph-Guided Scene Reconstruction from Images with 3D Gaussian Splatting by Chong Cheng Gaochao Song Yiyang Yao Qinzheng Zhou, Gangjian Zhang Hao Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-17377.txt.
Description
Chong Cheng and five co-authors present a study from twenty twenty-five on building large three dimensional scenes from plain photos. Past methods need exact camera poses or many views to learn from. Sparse views can also make the scene fit those views too well and look poor from new angles. Their method, called GraphGS, aims to fix these flaws.
First they guess the scene layout from space cues in the photos. That guess forms a graph of how the cameras link to each other. The graph then guides checks that views agree, and smart picks of which points to train on. Tests on many data sets show crisp, true scenes and state of the art scores. The team shares a project page with more detail.
Catalog record
| CREATOR | Chong Cheng Gaochao Song Yiyang Yao Qinzheng Zhou, Gangjian Zhang Hao Wang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-17377