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

Sequential manipulation tasks require a robot to perceive the state of anenvironment and plan a sequence of actions leading to a desired goal state. Insuch tasks, the ability to reason about spatial relations among object entitiesfrom raw sensor inputs is crucial in order to determine when a task has beencompleted and which actions can be executed. In this work, we propose SORNet(Spatial Object-Centric Representation Network), a framework for learningobject-centric representations from RGB images conditioned on a set of objectqueries, represented as image patches called canonical object views. With onlya single canonical view per object and no annotation, SORNet generalizeszero-shot to object entities whose shape and texture are both unseen duringtraining. We evaluate SORNet on various spatial reasoning tasks such as spatialrelation classification and relative direction regression in complex tabletopmanipulation scenarios and show that SORNet significantly outperforms baselinesincluding state-of-the-art representation learning techniques. We alsodemonstrate the application of the representation learned by SORNet onvisual-servoing and task planning for sequential manipulation on a real robot.

SORNet: Spatial Object-Centric Representations for Sequential Manipulation