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

We introduce MegaPose, a method to estimate the 6D pose of novel objects,that is, objects unseen during training. At inference time, the method onlyassumes knowledge of (i) a region of interest displaying the object in theimage and (ii) a CAD model of the observed object. The contributions of thiswork are threefold. First, we present a 6D pose refiner based on arender&compare strategy which can be applied to novel objects. The shape andcoordinate system of the novel object are provided as inputs to the network byrendering multiple synthetic views of the object's CAD model. Second, weintroduce a novel approach for coarse pose estimation which leverages a networktrained to classify whether the pose error between a synthetic rendering and anobserved image of the same object can be corrected by the refiner. Third, weintroduce a large-scale synthetic dataset of photorealistic images of thousandsof objects with diverse visual and shape properties and show that thisdiversity is crucial to obtain good generalization performance on novelobjects. We train our approach on this large synthetic dataset and apply itwithout retraining to hundreds of novel objects in real images from severalpose estimation benchmarks. Our approach achieves state-of-the-art performanceon the ModelNet and YCB-Video datasets. An extensive evaluation on the 7 coredatasets of the BOP challenge demonstrates that our approach achievesperformance competitive with existing approaches that require access to thetarget objects during training. Code, dataset and trained models are availableon the project page: https://megapose6d.github.io/.

MegaPose: 6D Pose Estimation of Novel Objects via Render & Compare