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

3D visual grounding is the task of localizing the object in a 3D scene whichis referred by a description in natural language. With a wide range ofapplications ranging from autonomous indoor robotics to AR/VR, the task hasrecently risen in popularity. A common formulation to tackle 3D visualgrounding is grounding-by-detection, where localization is done via boundingboxes. However, for real-life applications that require physical interactions,a bounding box insufficiently describes the geometry of an object. We thereforetackle the problem of dense 3D visual grounding, i.e. referral-based 3Dinstance segmentation. We propose a dense 3D grounding network ConcreteNet,featuring three novel stand-alone modules which aim to improve groundingperformance for challenging repetitive instances, i.e. instances withdistractors of the same semantic class. First, we introduce a bottom-upattentive fusion module that aims to disambiguate inter-instance relationalcues, next we construct a contrastive training scheme to induce separation inthe latent space, and finally we resolve view-dependent utterances via alearned global camera token. ConcreteNet ranks 1st on the challenging ScanReferonline benchmark by a considerable +9.43% accuracy at 50% IoU and has won theICCV 3rd Workshop on Language for 3D Scenes "3D Object Localization" challenge.

Three Ways to Improve Verbo-visual Fusion for Dense 3D Visual Grounding