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

Deep learning applications on LiDAR data suffer from a strong domain gap whenapplied to different sensors or tasks. In order for these methods to obtainsimilar accuracy on different data in comparison to values reported on publicbenchmarks, a large scale annotated dataset is necessary. However, in practicalapplications labeled data is costly and time consuming to obtain. Such factorshave triggered various research in label-efficient methods, but a large gapremains to their fully-supervised counterparts. Thus, we propose ImageTo360, aneffective and streamlined few-shot approach to label-efficient LiDARsegmentation. Our method utilizes an image teacher network to generate semanticpredictions for LiDAR data within a single camera view. The teacher is used topretrain the LiDAR segmentation student network, prior to optional fine-tuningon 360$^\circ$ data. Our method is implemented in a modular manner on the pointlevel and as such is generalizable to different architectures. We improve overthe current state-of-the-art results for label-efficient methods and evensurpass some traditional fully-supervised segmentation networks.

360$^\circ$ from a Single Camera: A Few-Shot Approach for LiDAR Segmentation