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

Recently, many 2D pretrained foundational models have demonstrated impressivezero-shot prediction capabilities. In this work, we design a novel pipeline forzero-shot 3D part segmentation, called ZeroPS. It high-quality transfersknowledge from 2D pretrained foundational models to 3D point clouds. The mainidea of our approach is to explore the natural relationship between multi-viewcorrespondences and the prompt mechanism of foundational models and buildbridges on it. Our pipeline consists of two components: 1) a self-extensioncomponent that extends 2D groups from a single viewpoint to spatialglobal-level 3D groups; 2) a multi-modal labeling component that introduces atwo-dimensional checking mechanism to vote each 2D predicted bounding box tothe best matching 3D part, and a Class Non-highest Vote Penalty function torefine the Vote Matrix. Additionally, a merging algorithm is included to mergepart-level 3D groups. Extensive evaluation of three zero-shot segmentationtasks on PartnetE datasets, achieving state-of-the-art results with significantimprovements (+19.6%, +5.2% and +4.9%, respectively) over existing methods. Ourproposed approach does not need any training, fine-tuning or learnableparameters. It is hardly affected by domain shift. The code will be released.

ZeroPS: High-quality Cross-modal Knowledge Transfer for Zero-Shot 3D Part Segmentation