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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Point cloud-based open-vocabulary 3D object detection aims to detect 3Dcategories that do not have ground-truth annotations in the training set. It isextremely challenging because of the limited data and annotations (boundingboxes with class labels or text descriptions) of 3D scenes. Previous approachesleverage large-scale richly-annotated image datasets as a bridge between 3D andcategory semantics but require an extra alignment process between 2D images and3D points, limiting the open-vocabulary ability of 3D detectors. Instead ofleveraging 2D images, we propose Object2Scene, the first approach thatleverages large-scale large-vocabulary 3D object datasets to augment existing3D scene datasets for open-vocabulary 3D object detection. Object2Scene insertsobjects from different sources into 3D scenes to enrich the vocabulary of 3Dscene datasets and generates text descriptions for the newly inserted objects.We further introduce a framework that unifies 3D detection and visualgrounding, named L3Det, and propose a cross-domain category-level contrastivelearning approach to mitigate the domain gap between 3D objects from differentdatasets. Extensive experiments on existing open-vocabulary 3D object detectionbenchmarks show that Object2Scene obtains superior performance over existingmethods. We further verify the effectiveness of Object2Scene on a new benchmarkOV-ScanNet-200, by holding out all rare categories as novel categories not seenduring training.

Object2Scene: Putting Objects in Context for Open-Vocabulary 3D Detection