UnLoc: Leveraging Depth Uncertainties for Floorplan Localization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wüest, Matthias, Engelmann, Francis, Miksik, Ondrej, Pollefeys, Marc, Barath, Daniel
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911516716957696
author Wüest, Matthias
Engelmann, Francis
Miksik, Ondrej
Pollefeys, Marc
Barath, Daniel
author_facet Wüest, Matthias
Engelmann, Francis
Miksik, Ondrej
Pollefeys, Marc
Barath, Daniel
contents We propose UnLoc, an efficient data-driven solution for sequential camera localization within floorplans. Floorplan data is readily available, long-term persistent, and robust to changes in visual appearance. We address key limitations of recent methods, such as the lack of uncertainty modeling in depth predictions and the necessity for custom depth networks trained for each environment. We introduce a novel probabilistic model that incorporates uncertainty estimation, modeling depth predictions as explicit probability distributions. By leveraging off-the-shelf pre-trained monocular depth models, we eliminate the need to rely on per-environment-trained depth networks, enhancing generalization to unseen spaces. We evaluate UnLoc on large-scale synthetic and real-world datasets, demonstrating significant improvements over existing methods in terms of accuracy and robustness. Notably, we achieve $2.7$ times higher localization recall on long sequences (100 frames) and $42.2$ times higher on short ones (15 frames) than the state of the art on the challenging LaMAR HGE dataset.
format Preprint
id arxiv_https___arxiv_org_abs_2509_11301
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle UnLoc: Leveraging Depth Uncertainties for Floorplan Localization
Wüest, Matthias
Engelmann, Francis
Miksik, Ondrej
Pollefeys, Marc
Barath, Daniel
Computer Vision and Pattern Recognition
We propose UnLoc, an efficient data-driven solution for sequential camera localization within floorplans. Floorplan data is readily available, long-term persistent, and robust to changes in visual appearance. We address key limitations of recent methods, such as the lack of uncertainty modeling in depth predictions and the necessity for custom depth networks trained for each environment. We introduce a novel probabilistic model that incorporates uncertainty estimation, modeling depth predictions as explicit probability distributions. By leveraging off-the-shelf pre-trained monocular depth models, we eliminate the need to rely on per-environment-trained depth networks, enhancing generalization to unseen spaces. We evaluate UnLoc on large-scale synthetic and real-world datasets, demonstrating significant improvements over existing methods in terms of accuracy and robustness. Notably, we achieve $2.7$ times higher localization recall on long sequences (100 frames) and $42.2$ times higher on short ones (15 frames) than the state of the art on the challenging LaMAR HGE dataset.
title UnLoc: Leveraging Depth Uncertainties for Floorplan Localization
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.11301