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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

6D object pose estimation is a crucial prerequisite for autonomous robotmanipulation applications. The state-of-the-art models for pose estimation areconvolutional neural network (CNN)-based. Lately, Transformers, an architectureoriginally proposed for natural language processing, is achievingstate-of-the-art results in many computer vision tasks as well. Equipped withthe multi-head self-attention mechanism, Transformers enable simplesingle-stage end-to-end architectures for learning object detection and 6Dobject pose estimation jointly. In this work, we propose YOLOPose (short formfor You Only Look Once Pose estimation), a Transformer-based multi-object 6Dpose estimation method based on keypoint regression and an improved variant ofthe YOLOPose model. In contrast to the standard heatmaps for predictingkeypoints in an image, we directly regress the keypoints. Additionally, weemploy a learnable orientation estimation module to predict the orientationfrom the keypoints. Along with a separate translation estimation module, ourmodel is end-to-end differentiable. Our method is suitable for real-timeapplications and achieves results comparable to state-of-the-art methods. Weanalyze the role of object queries in our architecture and reveal that theobject queries specialize in detecting objects in specific image regions.Furthermore, we quantify the accuracy trade-off of using datasets of smallersizes to train our model.

YOLOPose V2: Understanding and Improving Transformer-based 6D Pose Estimation