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

As RGB-D sensors become more affordable, using RGB-D images to obtainhigh-accuracy 6D pose estimation results becomes a better option.State-of-the-art approaches typically use different backbones to extractfeatures for RGB and depth images. They use a 2D CNN for RGB images and aper-pixel point cloud network for depth data, as well as a fusion network forfeature fusion. We find that the essential reason for using two independentbackbones is the "projection breakdown" problem. In the depth image plane, theprojected 3D structure of the physical world is preserved by the 1D depth valueand its built-in 2D pixel coordinate (UV). Any spatial transformation thatmodifies UV, such as resize, flip, crop, or pooling operations in the CNNpipeline, breaks the binding between the pixel value and UV coordinate. As aconsequence, the 3D structure is no longer preserved by a modified depth imageor feature. To address this issue, we propose a simple yet effective methoddenoted as Uni6D that explicitly takes the extra UV data along with RGB-Dimages as input. Our method has a Unified CNN framework for 6D pose estimationwith a single CNN backbone. In particular, the architecture of our method isbased on Mask R-CNN with two extra heads, one named RT head for directlypredicting 6D pose and the other named abc head for guiding the network to mapthe visible points to their coordinates in the 3D model as an auxiliary module.This end-to-end approach balances simplicity and accuracy, achieving comparableaccuracy with state of the arts and 7.2x faster inference speed on theYCB-Video dataset.

Uni6D: A Unified CNN Framework without Projection Breakdown for 6D Pose Estimation