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 (TAL) is a task of identifying a set of actionsin a video, which involves localizing the start and end frames and classifyingeach action instance. Existing methods have addressed this task by usingpredefined anchor windows or heuristic bottom-up boundary-matching strategies,which are major bottlenecks in inference time. Additionally, the main challengeis the inability to capture long-range actions due to a lack of globalcontextual information. In this paper, we present a novel anchor-freeframework, referred to as HTNet, which predicts a set of <start time, end time,class> triplets from a video based on a Transformer architecture. After theprediction of coarse boundaries, we refine it through a background featuresampling (BFS) module and hierarchical Transformers, which enables our model toaggregate global contextual information and effectively exploit the inherentsemantic relationships in a video. We demonstrate how our method localizesaccurate action instances and achieves state-of-the-art performance on two TALbenchmark datasets: THUMOS14 and ActivityNet 1.3.

HTNet: Anchor-free Temporal Action Localization with Hierarchical Transformers