PGGT-Loc: Primitive-Grounded Geometry Transformer for Feed-Forward Camera Localization
PGGT-Loc: Primitive-Grounded Geometry Transformer for Feed-Forward Camera Localization
This research introduces PGGT-Loc, a network designed for feed-forward 6-DoF camera localization by grounding visual observations to planar map primitives. The method aims to overcome the modality gap between map primitives and visual data, offering a scalable architecture that generalizes to unseen environments while achieving state-of-the-art performance on both dense surface maps and minimal layout models. Code and models are publicly available.
The paper proposes PGGT-Loc, a feed-forward network for 6-DoF camera localization that grounds query images to planar map primitives, aiming to reduce reliance on complex multi-stage pipelines and dedicated maps.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
- weights: not_reported
- dataset: not_reported
- code: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
https://github.com/3dv-casia/PGGT-Loc
Open source S5
Source：arXiv Robotics — research abstracts · arxiv.org