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

Temporal action localization in videos presents significant challenges in thefield of computer vision. While the boundary-sensitive method has been widelyadopted, its limitations include incomplete use of intermediate and globalinformation, as well as an inefficient proposal feature generator. To addressthese challenges, we propose a novel framework, Sparse Multilevel BoundaryGenerator (SMBG), which enhances the boundary-sensitive method with boundaryclassification and action completeness regression. SMBG features a multi-levelboundary module that enables faster processing by gathering boundaryinformation at different lengths. Additionally, we introduce a sparseextraction confidence head that distinguishes information inside and outsidethe action, further optimizing the proposal feature generator. To improve thesynergy between multiple branches and balance positive and negative samples, wepropose a global guidance loss. Our method is evaluated on two popularbenchmarks, ActivityNet-1.3 and THUMOS14, and is shown to achievestate-of-the-art performance, with a better inference speed (2.47xBSN++,2.12xDBG). These results demonstrate that SMBG provides a more efficient andsimple solution for generating temporal action proposals. Our proposedframework has the potential to advance the field of computer vision and enhancethe accuracy and speed of temporal action localization in video analysis.Thecode and models are made available at\url{https://github.com/zhouyang-001/SMBG-for-temporal-action-proposal}.

Faster Learning of Temporal Action Proposal via Sparse Multilevel Boundary Generator