Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network

Abstract Since the era of we‐media, live video industry has shown an explosive growth trend. For large‐scale live video streaming, especially those containing crowd events that may cause great social impact, how to identify and supervise the crowd activity in live video streaming effectively is of g...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Junpeng Kang, Jing Zhang, Wensheng Li, Li Zhuo
Formato:	article
Lenguaje:	EN
Publicado:	Wiley 2021
Materias:	Photography TR1-1050 Computer software QA76.75-76.765
Acceso en línea:	https://doaj.org/article/2c1e125c055446439dcfecf525920db4
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:2c1e125c055446439dcfecf525920db4
record_format	dspace
spelling	oai:doaj.org-article:2c1e125c055446439dcfecf525920db42021-11-29T03:38:16ZCrowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network1751-96671751-965910.1049/ipr2.12239https://doaj.org/article/2c1e125c055446439dcfecf525920db42021-12-01T00:00:00Zhttps://doi.org/10.1049/ipr2.12239https://doaj.org/toc/1751-9659https://doaj.org/toc/1751-9667Abstract Since the era of we‐media, live video industry has shown an explosive growth trend. For large‐scale live video streaming, especially those containing crowd events that may cause great social impact, how to identify and supervise the crowd activity in live video streaming effectively is of great value to push the healthy development of live video industry. The existing crowd activity recognition mainly uses visual information, rarely fully exploiting and utilizing the correlation or external knowledge between crowd content. Therefore, a crowd activity recognition method in live video streaming is proposed by 3D‐ResNet and regional graph convolution network (ReGCN). (1) After extracting deep spatiotemporal features from live video streaming with 3D‐ResNet, the region proposals are generated by region proposal network. (2) A weakly supervised ReGCN is constructed by making region proposals as graph nodes and their correlations as edges. (3) Crowd activity in live video streaming is recognised by combining the output of ReGCN, the deep spatiotemporal features and the crowd motion intensity as external knowledge. Four experiments are conducted on the public collective activity extended dataset and a real‐world dataset BJUT‐CAD. The competitive results demonstrate that our method can effectively recognise crowd activity in live video streaming.Junpeng KangJing ZhangWensheng LiLi ZhuoWileyarticlePhotographyTR1-1050Computer softwareQA76.75-76.765ENIET Image Processing, Vol 15, Iss 14, Pp 3476-3486 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Photography TR1-1050 Computer software QA76.75-76.765
spellingShingle	Photography TR1-1050 Computer software QA76.75-76.765 Junpeng Kang Jing Zhang Wensheng Li Li Zhuo Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network
description	Abstract Since the era of we‐media, live video industry has shown an explosive growth trend. For large‐scale live video streaming, especially those containing crowd events that may cause great social impact, how to identify and supervise the crowd activity in live video streaming effectively is of great value to push the healthy development of live video industry. The existing crowd activity recognition mainly uses visual information, rarely fully exploiting and utilizing the correlation or external knowledge between crowd content. Therefore, a crowd activity recognition method in live video streaming is proposed by 3D‐ResNet and regional graph convolution network (ReGCN). (1) After extracting deep spatiotemporal features from live video streaming with 3D‐ResNet, the region proposals are generated by region proposal network. (2) A weakly supervised ReGCN is constructed by making region proposals as graph nodes and their correlations as edges. (3) Crowd activity in live video streaming is recognised by combining the output of ReGCN, the deep spatiotemporal features and the crowd motion intensity as external knowledge. Four experiments are conducted on the public collective activity extended dataset and a real‐world dataset BJUT‐CAD. The competitive results demonstrate that our method can effectively recognise crowd activity in live video streaming.
format	article
author	Junpeng Kang Jing Zhang Wensheng Li Li Zhuo
author_facet	Junpeng Kang Jing Zhang Wensheng Li Li Zhuo
author_sort	Junpeng Kang
title	Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network
title_short	Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network
title_full	Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network
title_fullStr	Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network
title_full_unstemmed	Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network
title_sort	crowd activity recognition in live video streaming via 3d‐resnet and region graph convolution network
publisher	Wiley
publishDate	2021
url	https://doaj.org/article/2c1e125c055446439dcfecf525920db4
work_keys_str_mv	AT junpengkang crowdactivityrecognitioninlivevideostreamingvia3dresnetandregiongraphconvolutionnetwork AT jingzhang crowdactivityrecognitioninlivevideostreamingvia3dresnetandregiongraphconvolutionnetwork AT wenshengli crowdactivityrecognitioninlivevideostreamingvia3dresnetandregiongraphconvolutionnetwork AT lizhuo crowdactivityrecognitioninlivevideostreamingvia3dresnetandregiongraphconvolutionnetwork
_version_	1718407634964447232

Crowd activity recognition in live video streaming via 3D‐ResNet and region graph convolution network

Ejemplares similares