English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

LiDAR place recognition approaches based on deep learning suffer fromsignificant performance degradation when there is a shift between thedistribution of training and test datasets, often requiring re-training thenetworks to achieve peak performance. However, obtaining accurate ground truthdata for new training data can be prohibitively expensive, especially incomplex or GPS-deprived environments. To address this issue we proposeGeoAdapt, which introduces a novel auxiliary classification head to generatepseudo-labels for re-training on unseen environments in a self-supervisedmanner. GeoAdapt uses geometric consistency as a prior to improve therobustness of our generated pseudo-labels against domain shift, improving theperformance and reliability of our Test-Time Adaptation approach. Comprehensiveexperiments show that GeoAdapt significantly boosts place recognitionperformance across moderate to severe domain shifts, and is competitive withfully supervised test-time adaptation approaches. Our code is available athttps://github.com/csiro-robotics/GeoAdapt.

GeoAdapt: Self-Supervised Test-Time Adaptation in LiDAR Place Recognition Using Geometric Priors