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/

Zendy Plus

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

Zendy Tools

Zendy Open

Human group detection, which splits crowd of people into groups, is animportant step for video-based human social activity analysis. The core ofhuman group detection is the human social relation representation anddivision.In this paper, we propose a new two-stage multi-head framework forhuman group detection. In the first stage, we propose a human behaviorsimulator head to learn the social relation feature embedding, which isself-supervisely trained by leveraging the socially grounded multi-personbehavior relationship. In the second stage, based on the social relationembedding, we develop a self-attention inspired network for human groupdetection. Remarkable performance on two state-of-the-art large-scalebenchmarks, i.e., PANDA and JRDB-Group, verifies the effectiveness of theproposed framework. Benefiting from the self-supervised social relationembedding, our method can provide promising results with very few (labeled)training data. We will release the source code to the public.

Self-supervised Social Relation Representation for Human Group Detection