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

Skeleton based action recognition distinguishes human actions using thetrajectories of skeleton joints, which provide a very good representation fordescribing actions. Considering that recurrent neural networks (RNNs) with LongShort-Term Memory (LSTM) can learn feature representations and model long-termtemporal dependencies automatically, we propose an end-to-end fully connecteddeep LSTM network for skeleton based action recognition. Inspired by theobservation that the co-occurrences of the joints intrinsically characterizehuman actions, we take the skeleton as the input at each time slot andintroduce a novel regularization scheme to learn the co-occurrence features ofskeleton joints. To train the deep LSTM network effectively, we propose a newdropout algorithm which simultaneously operates on the gates, cells, and outputresponses of the LSTM neurons. Experimental results on three human actionrecognition datasets consistently demonstrate the effectiveness of the proposedmodel.

Co-occurrence Feature Learning for Skeleton based Action Recognition using Regularized Deep LSTM Networks