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 video grounding (TVG) aims to retrieve the time interval of alanguage query from an untrimmed video. A significant challenge in TVG is thelow "Semantic Noise Ratio (SNR)", which results in worse performance with lowerSNR. Prior works have addressed this challenge using sophisticated techniques.In this paper, we propose a no-frills TVG model that consists of two coremodules, namely multi-scale neighboring attention and zoom-in boundarydetection. The multi-scale neighboring attention restricts each video token toonly aggregate visual contexts from its neighbor, enabling the extraction ofthe most distinguishing information with multi-scale feature hierarchies fromhigh-ratio noises. The zoom-in boundary detection then focuses on local-wisediscrimination of the selected top candidates for fine-grained groundingadjustment. With an end-to-end training strategy, our model achievescompetitive performance on different TVG benchmarks, while also having theadvantage of faster inference speed and lighter model parameters, thanks to itslightweight architecture.

No-frills Temporal Video Grounding: Multi-Scale Neighboring Attention and Zoom-in Boundary Detection