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 Plus annual

Presents the access of premium content as premium feature

Premium Content

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 Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Localization is a critically essential and crucial enabler of autonomousrobots. While deep learning has made significant strides in many computervision tasks, it is still yet to make a sizeable impact on improvingcapabilities of metric visual localization. One of the major hindrances hasbeen the inability of existing Convolutional Neural Network (CNN)-based poseregression methods to generalize to previously unseen places. Our recentlyintroduced CMRNet effectively addresses this limitation by enabling mapindependent monocular localization in LiDAR-maps. In this paper, we now take ita step further by introducing CMRNet++, which is a significantly more robustmodel that not only generalizes to new places effectively, but is alsoindependent of the camera parameters. We enable this capability by combiningdeep learning with geometric techniques, and by moving the metric reasoningoutside the learning process. In this way, the weights of the network are nottied to a specific camera. Extensive evaluations of CMRNet++ on threechallenging autonomous driving datasets, i.e., KITTI, Argoverse, and Lyft5,show that CMRNet++ outperforms CMRNet as well as other baselines by a largemargin. More importantly, for the first-time, we demonstrate the ability of adeep learning approach to accurately localize without any retraining orfine-tuning in a completely new environment and independent of the cameraparameters.

CMRNet++: Map and Camera Agnostic Monocular Visual Localization in LiDAR Maps