GigaSLAM: Large-Scale Monocular SLAM with Hierarchical Gaussian Splats
Kai Deng, Yigong Zhang, Jian Yang, Jin Xie · 2025
Read GigaSLAM: Large-Scale Monocular SLAM with Hierarchical Gaussian Splats by Kai Deng, Yigong Zhang, Jian Yang, Jin Xie (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-2503-08071.txt.
Description
Kai Deng, Yigong Zhang, Jian Yang, and Jin Xie stretch camera-only mapping to kilometer scale. Simultaneous localization and mapping with one camera fails outdoors in most neural systems, which were built for single rooms. Their 2025 framework, GigaSLAM, stores the world as hierarchical sparse voxels. Neural decoders render Gaussian splats at several detail levels from that map. A metric depth model, epipolar geometry, and classic pose solving drive the tracking front end.
A bag-of-words loop closer keeps long drives aligned. Four driving datasets, KITTI among them, test the system. Tracking stays precise and renderings faithful across whole city stretches. The team claims the first Gaussian splatting SLAM proven at this scale. Code is public for the self-driving and robotics crowds.
Catalog record
| CREATOR | Kai Deng, Yigong Zhang, Jian Yang, Jin Xie |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-08071