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

Learning dense point-wise semantics from unstructured 3D point clouds withfewer labels, although a realistic problem, has been under-explored inliterature. While existing weakly supervised methods can effectively learnsemantics with only a small fraction of point-level annotations, we find thatthe vanilla bounding box-level annotation is also informative for semanticsegmentation of large-scale 3D point clouds. In this paper, we introduce aneural architecture, termed Box2Seg, to learn point-level semantics of 3D pointclouds with bounding box-level supervision. The key to our approach is togenerate accurate pseudo labels by exploring the geometric and topologicalstructure inside and outside each bounding box. Specifically, anattention-based self-training (AST) technique and Point Class ActivationMapping (PCAM) are utilized to estimate pseudo-labels. The network is furthertrained and refined with pseudo labels. Experiments on two large-scalebenchmarks including S3DIS and ScanNet demonstrate the competitive performanceof the proposed method. In particular, the proposed network can be trained withcheap, or even off-the-shelf bounding box-level annotations and subcloud-leveltags.

Box2Seg: Learning Semantics of 3D Point Clouds with Box-Level Supervision