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

Real-scanned point clouds are often incomplete due to viewpoint, occlusion,and noise, which hampers 3D geometric modeling and perception. Existing pointcloud completion methods tend to generate global shape skeletons and hence lackfine local details. Furthermore, they mostly learn a deterministicpartial-to-complete mapping, but overlook structural relations in man-madeobjects. To tackle these challenges, this paper proposes a variationalframework, Variational Relational point Completion Network (VRCNet) with twoappealing properties: 1) Probabilistic Modeling. In particular, we propose adual-path architecture to enable principled probabilistic modeling acrosspartial and complete clouds. One path consumes complete point clouds forreconstruction by learning a point VAE. The other path generates completeshapes for partial point clouds, whose embedded distribution is guided bydistribution obtained from the reconstruction path during training. 2)Relational Enhancement. Specifically, we carefully design point self-attentionkernel and point selective kernel module to exploit relational point features,which refines local shape details conditioned on the coarse completion. Inaddition, we contribute multi-view partial point cloud datasets (MVP and MVP-40dataset) containing over 200,000 high-quality scans, which render partial 3Dshapes from 26 uniformly distributed camera poses for each 3D CAD model.Extensive experiments demonstrate that VRCNet outperforms state-of-the-artmethods on all standard point cloud completion benchmarks. Notably, VRCNetshows great generalizability and robustness on real-world point cloud scans.Moreover, we can achieve robust 3D classification for partial point clouds withthe help of VRCNet, which can highly increase classification accuracy.

Variational Relational Point Completion Network for Robust 3D Classification