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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Temporal action recognition always depends on temporal action proposalgeneration to hypothesize actions and algorithms usually need to process verylong video sequences and output the starting and ending times of each potentialaction in each video suffering from high computation cost. To address this,based on boundary sensitive network we propose a new temporal convolutionnetwork called Multipath Temporal ConvNet (MTN), which consists of two partsi.e. Multipath DenseNet and SE-ConvNet. In this work, one novel highperformance ring parallel architecture based on Message Passing Interface (MPI)is further introduced into temporal action proposal generation, which is areliable communication protocol, in order to respond to the requirements oflarge memory occupation and a large number of videos. Remarkably, the totaldata transmission is reduced by adding a connection between multiple computingload in the newly developed architecture. It is found that, compared to thetraditional Parameter Server architecture, our parallel architecture has higherefficiency on temporal action detection task with multiple GPUs, which issuitable for dealing with the tasks of temporal action proposal generation,especially for large datasets of millions of videos. We conduct experiments onActivityNet-1.3 and THUMOS14, where our method outperforms other state-of-arttemporal action detection methods with high recall and high temporal precision.In addition, a time metric is further proposed here to evaluate the speedperformance in the distributed training process.

Accelerating temporal action proposal generation via high performance computing