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

The task of 3D semantic scene graph (3DSSG) prediction in the point cloud ischallenging since (1) the 3D point cloud only captures geometric structureswith limited semantics compared to 2D images, and (2) long-tailed relationdistribution inherently hinders the learning of unbiased prediction. Since 2Dimages provide rich semantics and scene graphs are in nature coped withlanguages, in this study, we propose Visual-Linguistic Semantics AssistedTraining (VL-SAT) scheme that can significantly empower 3DSSG prediction modelswith discrimination about long-tailed and ambiguous semantic relations. The keyidea is to train a powerful multi-modal oracle model to assist the 3D model.This oracle learns reliable structural representations based on semantics fromvision, language, and 3D geometry, and its benefits can be heterogeneouslypassed to the 3D model during the training stage. By effectively utilizingvisual-linguistic semantics in training, our VL-SAT can significantly boostcommon 3DSSG prediction models, such as SGFN and SGGpoint, only with 3D inputsin the inference stage, especially when dealing with tail relation triplets.Comprehensive evaluations and ablation studies on the 3DSSG dataset havevalidated the effectiveness of the proposed scheme. Code is available athttps://github.com/wz7in/CVPR2023-VLSAT.

VL-SAT: Visual-Linguistic Semantics Assisted Training for 3D Semantic Scene Graph Prediction in Point Cloud