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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

This paper presents 6D-ViT, a transformer-based instance representationlearning network, which is suitable for highly accurate category-level objectpose estimation on RGB-D images. Specifically, a novel two-streamencoder-decoder framework is dedicated to exploring complex and powerfulinstance representations from RGB images, point clouds and categorical shapepriors. For this purpose, the whole framework consists of two main branches,named Pixelformer and Pointformer. The Pixelformer contains a pyramidtransformer encoder with an all-MLP decoder to extract pixelwise appearancerepresentations from RGB images, while the Pointformer relies on a cascadedtransformer encoder and an all-MLP decoder to acquire the pointwise geometriccharacteristics from point clouds. Then, dense instance representations (i.e.,correspondence matrix, deformation field) are obtained from a multi-sourceaggregation network with shape priors, appearance and geometric information asinput. Finally, the instance 6D pose is computed by leveraging thecorrespondence among dense representations, shape priors, and the instancepoint clouds. Extensive experiments on both synthetic and real-world datasetsdemonstrate that the proposed 3D instance representation learning frameworkachieves state-of-the-art performance on both datasets, and significantlyoutperforms all existing methods.

6D-ViT: Category-Level 6D Object Pose Estimation via Transformer-Based Instance Representation Learning