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

Insufficient labeled training datasets is one of the bottlenecks of 3D handpose estimation from monocular RGB images. Synthetic datasets have a largenumber of images with precise annotations, but the obvious difference withreal-world datasets impacts the generalization. Little work has been done tobridge the gap between two domains over their wide difference. In this paper,we propose a domain adaptation method called Adaptive Wasserstein Hourglass (AWHourglass) for weakly-supervised 3D hand pose estimation, which aims todistinguish the difference and explore the common characteristics (e.g. handstructure) of synthetic and real-world datasets. Learning the commoncharacteristics helps the network focus on pose-related information. Thesimilarity of the characteristics makes it easier to enforce domain-invariantconstraints. During training, based on the relation between these commoncharacteristics and 3D pose learned from fully-annotated synthetic datasets, itis beneficial for the network to restore the 3D pose of weakly labeledreal-world datasets with the aid of 2D annotations and depth images. While intesting, the network predicts the 3D pose with the input of RGB.

Adaptive Wasserstein Hourglass for Weakly Supervised Hand Pose Estimation from Monocular RGB