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

The point process is a solid framework to model sequential data, such asvideos, by exploring the underlying relevance. As a challenging problem forhigh-level video understanding, weakly supervised action recognition andlocalization in untrimmed videos has attracted intensive research attention.Knowledge transfer by leveraging the publicly available trimmed videos asexternal guidance is a promising attempt to make up for the coarse-grainedvideo-level annotation and improve the generalization performance. However,unconstrained knowledge transfer may bring about irrelevant noise andjeopardize the learning model. This paper proposes a novel adaptabilitydecomposing encoder-decoder network to transfer reliable knowledge betweentrimmed and untrimmed videos for action recognition and localization viabidirectional point process modeling, given only video-level annotations. Bydecomposing the original features into domain-adaptable and domain-specificones based on their adaptability, trimmed-untrimmed knowledge transfer can besafely confined within a more coherent subspace. An encoder-decoder basedstructure is carefully designed and jointly optimized to facilitate effectiveaction classification and temporal localization. Extensive experiments areconducted on two benchmark datasets (i.e., THUMOS14 and ActivityNet1.3), andexperimental results clearly corroborate the efficacy of our method.

AdapNet: Adaptability Decomposing Encoder-Decoder Network for Weakly Supervised Action Recognition and Localization