Density-Aware Multi-Task Learning for Crowd Counting

doi:10.1109/TMM.2020.2980945

CORC > 自动化研究所 > 中国科学院自动化研究所 > 模式识别国家重点实验室 > 多媒体计算与图形学团队

	Density-Aware Multi-Task Learning for Crowd Counting
	Jiang, Xiaoheng 2; Zhang, Li 2; Zhang, Tianzhu 1; Lv, Pei 2; Zhou, Bing 2; Pang, Yanwei 4; Xu, Mingliang 2; Xu, Changsheng 3
刊名	IEEE TRANSACTIONS ON MULTIMEDIA
	2021
卷号	23 页码:443-453
关键词	Task analysis Semantics Estimation Feature extraction Convolutional neural networks Cameras Head Convolutional neural network crowd counting density-level classification density map estimation multi-task learning
ISSN号	1520-9210
DOI	10.1109/TMM.2020.2980945
通讯作者	Xu, Mingliang(iexumingliang@zzu.edu.cn)
英文摘要	In this paper, we present a method called density-aware convolutional neural network (DensityCNN) to perform the crowd counting task in various crowded scenes. The key idea of the DensityCNN is to utilize high-level semantic information to provide guidance and constraint when generating density maps. To this end, we implement the DensityCNN by adopting a multi-task CNN structure to jointly learn density-level classification and density map estimation. The density-level classification task learns multi-channel semantic features that are aware of the density distributions of the input image. This task is accomplished via our specially designed group-based convolutional structure in a supervised learning manner. In the density map estimation task, these semantic features are deployed together with high-dimension convolutional features to generate density maps with lower count errors. Extensive experiments on four challenging crowd datasets (ShanghaiTech, UCF_CC_50, UCF-QNCF, and WorldExpo'10) and one vehicle dataset TRANCOS demonstrate the effectiveness of the proposed method.
资助项目	National Natural Science Foundation of China[61802351] ; National Natural Science Foundation of China[61822701] ; National Natural Science Foundation of China[61872324] ; National Natural Science Foundation of China[61772474] ; China Postdoctoral Science Foundation[2018M632802] ; Key R&D and Promotion Projects in Henan Province[192102310258]
WOS关键词	DEEP
WOS研究方向	Computer Science ; Telecommunications
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000601877600035
资助机构	National Natural Science Foundation of China ; China Postdoctoral Science Foundation ; Key R&D and Promotion Projects in Henan Province
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/42826]
专题	自动化研究所_模式识别国家重点实验室_多媒体计算与图形学团队
通讯作者	Xu, Mingliang
作者单位	1.Univ Sci & Technol China, Hefei 230026, Peoples R China 2.Zhengzhou Univ, Sch Informat Engn, Zhengzhou 450001, Peoples R China 3.Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China 4.Tianjin Univ, Sch Elect & Informat Engn, Tianjin 300072, Peoples R China
推荐引用方式 GB/T 7714	Jiang, Xiaoheng,Zhang, Li,Zhang, Tianzhu,et al. Density-Aware Multi-Task Learning for Crowd Counting[J]. IEEE TRANSACTIONS ON MULTIMEDIA,2021,23:443-453.
APA	Jiang, Xiaoheng.,Zhang, Li.,Zhang, Tianzhu.,Lv, Pei.,Zhou, Bing.,...&Xu, Changsheng.(2021).Density-Aware Multi-Task Learning for Crowd Counting.IEEE TRANSACTIONS ON MULTIMEDIA,23,443-453.
MLA	Jiang, Xiaoheng,et al."Density-Aware Multi-Task Learning for Crowd Counting".IEEE TRANSACTIONS ON MULTIMEDIA 23(2021):443-453.