Behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations

Abstract Detecting human behaviours in images of crowded classroom scenes is a challenging task, due to the large variations of humans in scale and pose perspective. In this paper, two modules are proposed to tackle these two variations. First, an attention‐based RoI (region‐of‐interest) extractor i...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Mingyu Liu, Fanman Meng, Qingbo Wu, Linfeng Xu, Qianghua Liao
Formato:	article
Lenguaje:	EN
Publicado:	Wiley 2021
Materias:	Photography TR1-1050 Computer software QA76.75-76.765
Acceso en línea:	https://doaj.org/article/e18bac49cba44e608faa4e3b7caa524d
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:e18bac49cba44e608faa4e3b7caa524d
record_format	dspace
spelling	oai:doaj.org-article:e18bac49cba44e608faa4e3b7caa524d2021-11-29T03:38:16ZBehaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations1751-96671751-965910.1049/ipr2.12318https://doaj.org/article/e18bac49cba44e608faa4e3b7caa524d2021-12-01T00:00:00Zhttps://doi.org/10.1049/ipr2.12318https://doaj.org/toc/1751-9659https://doaj.org/toc/1751-9667Abstract Detecting human behaviours in images of crowded classroom scenes is a challenging task, due to the large variations of humans in scale and pose perspective. In this paper, two modules are proposed to tackle these two variations. First, an attention‐based RoI (region‐of‐interest) extractor is designed to handle scale variation. Feature fusion and attention mechanism are used to improve the RoI feature with more local and global information. Second, a transformation‐based detection head is introduced to handle perspective variation. The spatial transformation is adopted to extract consistent representation under various perspectives. Moreover, since there is a lack of proper datasets for human behaviour detection in classroom scenes, a new dataset is created, namely CLBD. The experiments on the proposed dataset demonstrate that the modules obtain significant improvements of performance over the state‐of‐the‐art detectors.Mingyu LiuFanman MengQingbo WuLinfeng XuQianghua LiaoWileyarticlePhotographyTR1-1050Computer softwareQA76.75-76.765ENIET Image Processing, Vol 15, Iss 14, Pp 3466-3475 (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 Mingyu Liu Fanman Meng Qingbo Wu Linfeng Xu Qianghua Liao Behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations
description	Abstract Detecting human behaviours in images of crowded classroom scenes is a challenging task, due to the large variations of humans in scale and pose perspective. In this paper, two modules are proposed to tackle these two variations. First, an attention‐based RoI (region‐of‐interest) extractor is designed to handle scale variation. Feature fusion and attention mechanism are used to improve the RoI feature with more local and global information. Second, a transformation‐based detection head is introduced to handle perspective variation. The spatial transformation is adopted to extract consistent representation under various perspectives. Moreover, since there is a lack of proper datasets for human behaviour detection in classroom scenes, a new dataset is created, namely CLBD. The experiments on the proposed dataset demonstrate that the modules obtain significant improvements of performance over the state‐of‐the‐art detectors.
format	article
author	Mingyu Liu Fanman Meng Qingbo Wu Linfeng Xu Qianghua Liao
author_facet	Mingyu Liu Fanman Meng Qingbo Wu Linfeng Xu Qianghua Liao
author_sort	Mingyu Liu
title	Behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations
title_short	Behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations
title_full	Behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations
title_fullStr	Behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations
title_full_unstemmed	Behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations
title_sort	behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations
publisher	Wiley
publishDate	2021
url	https://doaj.org/article/e18bac49cba44e608faa4e3b7caa524d
work_keys_str_mv	AT mingyuliu behaviourdetectionincrowdedclassroomscenesviaenhancingfeaturesrobusttoscaleandperspectivevariations AT fanmanmeng behaviourdetectionincrowdedclassroomscenesviaenhancingfeaturesrobusttoscaleandperspectivevariations AT qingbowu behaviourdetectionincrowdedclassroomscenesviaenhancingfeaturesrobusttoscaleandperspectivevariations AT linfengxu behaviourdetectionincrowdedclassroomscenesviaenhancingfeaturesrobusttoscaleandperspectivevariations AT qianghualiao behaviourdetectionincrowdedclassroomscenesviaenhancingfeaturesrobusttoscaleandperspectivevariations
_version_	1718407679148294144

Behaviour detection in crowded classroom scenes via enhancing features robust to scale and perspective variations

Ejemplares similares