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

Although many approaches for multi-human pose estimation in videos have shownprofound results, they require densely annotated data which entails excessiveman labor. Furthermore, there exists occlusion and motion blur that inevitablylead to poor estimation performance. To address these problems, we propose amethod that leverages an attention mask for occluded joints and encodestemporal dependency between frames using transformers. First, our frameworkcomposes different combinations of sparsely annotated frames that denote thetrack of the overall joint movement. We propose an occlusion attention maskfrom these combinations that enable encoding occlusion-aware heatmaps as asemi-supervised task. Second, the proposed temporal encoder employs transformerarchitecture to effectively aggregate the temporal relationship andkeypoint-wise attention from each time step and accurately refines the targetframe's final pose estimation. We achieve state-of-the-art pose estimationresults for PoseTrack2017 and PoseTrack2018 datasets and demonstrate therobustness of our approach to occlusion and motion blur in sparsely annotatedvideo data.

OTPose: Occlusion-Aware Transformer for Pose Estimation in Sparsely-Labeled Videos