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

In this paper, we study the problem of 3D object segmentation from raw pointclouds. Unlike all existing methods which usually require a large amount ofhuman annotations for full supervision, we propose the first unsupervisedmethod, called OGC, to simultaneously identify multiple 3D objects in a singleforward pass, without needing any type of human annotations. The key to ourapproach is to fully leverage the dynamic motion patterns over sequential pointclouds as supervision signals to automatically discover rigid objects. Ourmethod consists of three major components, 1) the object segmentation networkto directly estimate multi-object masks from a single point cloud frame, 2) theauxiliary self-supervised scene flow estimator, and 3) our core object geometryconsistency component. By carefully designing a series of loss functions, weeffectively take into account the multi-object rigid consistency and the objectshape invariance in both temporal and spatial scales. This allows our method totruly discover the object geometry even in the absence of annotations. Weextensively evaluate our method on five datasets, demonstrating the superiorperformance for object part instance segmentation and general objectsegmentation in both indoor and the challenging outdoor scenarios.

OGC: Unsupervised 3D Object Segmentation from Rigid Dynamics of Point Clouds