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 address the highly challenging problem of real-time 3D hand tracking basedon a monocular RGB-only sequence. Our tracking method combines a convolutionalneural network with a kinematic 3D hand model, such that it generalizes well tounseen data, is robust to occlusions and varying camera viewpoints, and leadsto anatomically plausible as well as temporally smooth hand motions. Fortraining our CNN we propose a novel approach for the synthetic generation oftraining data that is based on a geometrically consistent image-to-imagetranslation network. To be more specific, we use a neural network thattranslates synthetic images to "real" images, such that the so-generated imagesfollow the same statistical distribution as real-world hand images. Fortraining this translation network we combine an adversarial loss and acycle-consistency loss with a geometric consistency loss in order to preservegeometric properties (such as hand pose) during translation. We demonstratethat our hand tracking system outperforms the current state-of-the-art onchallenging RGB-only footage.

GANerated Hands for Real-time 3D Hand Tracking from Monocular RGB