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

Most existing methods for category-level pose estimation rely on object pointclouds. However, when considering transparent objects, depth cameras areusually not able to capture meaningful data, resulting in point clouds withsevere artifacts. Without a high-quality point cloud, existing methods are notapplicable to challenging transparent objects. To tackle this problem, wepresent StereoPose, a novel stereo image framework for category-level objectpose estimation, ideally suited for transparent objects. For a robustestimation from pure stereo images, we develop a pipeline that decouplescategory-level pose estimation into object size estimation, initial poseestimation, and pose refinement. StereoPose then estimates object pose based onrepresentation in the normalized object coordinate space~(NOCS). To address theissue of image content aliasing, we further define a back-view NOCS map for thetransparent object. The back-view NOCS aims to reduce the network learningambiguity caused by content aliasing, and leverage informative cues on the backof the transparent object for more accurate pose estimation. To further improvethe performance of the stereo framework, StereoPose is equipped with a parallaxattention module for stereo feature fusion and an epipolar loss for improvingthe stereo-view consistency of network predictions. Extensive experiments onthe public TOD dataset demonstrate the superiority of the proposed StereoPoseframework for category-level 6D transparent object pose estimation.

StereoPose: Category-Level 6D Transparent Object Pose Estimation from Stereo Images via Back-View NOCS