3D Gaussian Splatting aided Localization for Large and Complex Indoor-Environments
***** (for review, names must be rendered anonymous), Vincent Ress 1, Jonas Meyer 1,2, Wei Zhang 1, David Skuddis 1, Uwe Soergel 1, Norbert Haala 1 · 2025
Read 3D Gaussian Splatting aided Localization for Large and Complex Indoor-Environments by ***** (for review, names must be rendered anonymous), Vincent Ress 1, Jonas Meyer 1,2, Wei Zhang 1, David Skuddis 1, Uwe Soergel 1, Norbert Haala 1 (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-2502-13803.txt.
Description
Vincent Ress and five co-authors improve indoor visual localization with images rendered from a 3D Gaussian Splatting map, in work from 2025. Cameras place themselves by matching views against reference data, yet hard cases still break standard methods. Their pipeline first builds the splat map with a modern SLAM system, then renders extra reference views from poses sampled at random.
Both geometry-based localization and scene coordinate regression improved once the rendered views joined the reference set. Tests ran in a large industrial hall, the kind of space where plain methods falter. The authors analyse where the synthetic views help most, arguing that cheap rendered data can shore up real reference sets.
Catalog record
| CREATOR | ***** (for review, names must be rendered anonymous), Vincent Ress 1, Jonas Meyer 1,2, Wei Zhang 1, David Skuddis 1, Uwe Soergel 1, Norbert Haala 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-13803