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

Graph convolutional networks (GCNs) have been widely used and achievedremarkable results in skeleton-based action recognition. We think the key toskeleton-based action recognition is a skeleton hanging in frames, so we focuson how the Graph Convolutional Convolution networks learn different topologiesand effectively aggregate joint features in the global temporal and localtemporal. In this work, we propose three Channel-wise Tolopogy GraphConvolution based on Channel-wise Topology Refinement Graph Convolution(CTR-GCN). Combining CTR-GCN with two joint cross-attention modules can capturethe upper-lower body part and hand-foot relationship skeleton features. Afterthat, to capture features of human skeletons changing in frames we design theTemporal Attention Transformers to extract skeletons effectively. The TemporalAttention Transformers can learn the temporal features of human skeletonsequences. Finally, we fuse the temporal features output scale with MLP andclassification. We develop a powerful graph convolutional network named SpatialTemporal Effective Body-part Cross Attention Transformer which notablyhigh-performance on the NTU RGB+D, NTU RGB+D 120 datasets. Our code and modelsare available at https://github.com/maclong01/STEP-CATFormer

STEP CATFormer: Spatial-Temporal Effective Body-Part Cross Attention Transformer for Skeleton-based Action Recognition