Multiscale Aggregate Networks with Dense Connections for Crowd Counting

The most advanced method for crowd counting uses a fully convolutional network that extracts image features and then generates a crowd density map. However, this process often encounters multiscale and contextual loss problems. To address these problems, we propose a multiscale aggregation network (...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Pengfei Li, Min Zhang, Jian Wan, Ming Jiang
Formato:	article
Lenguaje:	EN
Publicado:	Hindawi Limited 2021
Materias:	Computer applications to medicine. Medical informatics R858-859.7 Neurosciences. Biological psychiatry. Neuropsychiatry RC321-571
Acceso en línea:	https://doaj.org/article/729f4a27483344b281121e3c1524c85f
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:729f4a27483344b281121e3c1524c85f
record_format	dspace
spelling	oai:doaj.org-article:729f4a27483344b281121e3c1524c85f2021-11-22T01:10:35ZMultiscale Aggregate Networks with Dense Connections for Crowd Counting1687-527310.1155/2021/9996232https://doaj.org/article/729f4a27483344b281121e3c1524c85f2021-01-01T00:00:00Zhttp://dx.doi.org/10.1155/2021/9996232https://doaj.org/toc/1687-5273The most advanced method for crowd counting uses a fully convolutional network that extracts image features and then generates a crowd density map. However, this process often encounters multiscale and contextual loss problems. To address these problems, we propose a multiscale aggregation network (MANet) that includes a feature extraction encoder (FEE) and a density map decoder (DMD). The FEE uses a cascaded scale pyramid network to extract multiscale features and obtains contextual features through dense connections. The DMD uses deconvolution and fusion operations to generate features containing detailed information. These features can be further converted into high-quality density maps to accurately calculate the number of people in a crowd. An empirical comparison using four mainstream datasets (ShanghaiTech, WorldExpo’10, UCF_CC_50, and SmartCity) shows that the proposed method is more effective in terms of the mean absolute error and mean squared error. The source code is available at https://github.com/lpfworld/MANet.Pengfei LiMin ZhangJian WanMing JiangHindawi LimitedarticleComputer applications to medicine. Medical informaticsR858-859.7Neurosciences. Biological psychiatry. NeuropsychiatryRC321-571ENComputational Intelligence and Neuroscience, Vol 2021 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Computer applications to medicine. Medical informatics R858-859.7 Neurosciences. Biological psychiatry. Neuropsychiatry RC321-571
spellingShingle	Computer applications to medicine. Medical informatics R858-859.7 Neurosciences. Biological psychiatry. Neuropsychiatry RC321-571 Pengfei Li Min Zhang Jian Wan Ming Jiang Multiscale Aggregate Networks with Dense Connections for Crowd Counting
description	The most advanced method for crowd counting uses a fully convolutional network that extracts image features and then generates a crowd density map. However, this process often encounters multiscale and contextual loss problems. To address these problems, we propose a multiscale aggregation network (MANet) that includes a feature extraction encoder (FEE) and a density map decoder (DMD). The FEE uses a cascaded scale pyramid network to extract multiscale features and obtains contextual features through dense connections. The DMD uses deconvolution and fusion operations to generate features containing detailed information. These features can be further converted into high-quality density maps to accurately calculate the number of people in a crowd. An empirical comparison using four mainstream datasets (ShanghaiTech, WorldExpo’10, UCF_CC_50, and SmartCity) shows that the proposed method is more effective in terms of the mean absolute error and mean squared error. The source code is available at https://github.com/lpfworld/MANet.
format	article
author	Pengfei Li Min Zhang Jian Wan Ming Jiang
author_facet	Pengfei Li Min Zhang Jian Wan Ming Jiang
author_sort	Pengfei Li
title	Multiscale Aggregate Networks with Dense Connections for Crowd Counting
title_short	Multiscale Aggregate Networks with Dense Connections for Crowd Counting
title_full	Multiscale Aggregate Networks with Dense Connections for Crowd Counting
title_fullStr	Multiscale Aggregate Networks with Dense Connections for Crowd Counting
title_full_unstemmed	Multiscale Aggregate Networks with Dense Connections for Crowd Counting
title_sort	multiscale aggregate networks with dense connections for crowd counting
publisher	Hindawi Limited
publishDate	2021
url	https://doaj.org/article/729f4a27483344b281121e3c1524c85f
work_keys_str_mv	AT pengfeili multiscaleaggregatenetworkswithdenseconnectionsforcrowdcounting AT minzhang multiscaleaggregatenetworkswithdenseconnectionsforcrowdcounting AT jianwan multiscaleaggregatenetworkswithdenseconnectionsforcrowdcounting AT mingjiang multiscaleaggregatenetworkswithdenseconnectionsforcrowdcounting
_version_	1718418345293774848

Multiscale Aggregate Networks with Dense Connections for Crowd Counting

Ejemplares similares