3D PostureNet: A unified framework for skeleton-based posture recognition

doi:10.1016/j.patrec.2020.09.029

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 > 遥感图像处理团队

	3D PostureNet: A unified framework for skeleton-based posture recognition
	Liu, Jianbo 1,2; Wang, Ying 1; Liu, Yongcheng 1,2; Xiang, Shiming 1; Pan, Chunhong 1
刊名	PATTERN RECOGNITION LETTERS
	2020-12-01
卷号	140 页码:143-149
关键词	Human posture recognition Static hand gesture recognition Skeleton-based 3D convolutional neural network
ISSN号	0167-8655
DOI	10.1016/j.patrec.2020.09.029
通讯作者	Wang, Ying(ywang@nlpr.ia.ac.cn)
英文摘要	Image-based posture recognition is a very challenging problem as it is difficult to acquire rich 3D information from postures in 2D images. Existing methods founded on 3D skeleton cues could alleviate this issue, but they are not particularly efficient due to the application of handcrafted features and traditional classifiers. This paper presents a novel and unified framework for skeleton-based posture recognition, applying powerful 3D Convolutional Neural Network (CNN) to this issue. Technically, bounding-box-based normalization for the raw skeleton data is proposed to eliminate the coordinate differences caused by diverse recording environments and posture displacements. Moreover, Gaussian voxelization for the skeleton is employed to expressively represent the posture configuration. Thereby, an end-to-end framework based on 3D CNN, called 3D PostureNet, is developed for robust posture recognition. To verify its effectiveness, a large-scale writing posture dataset is created and released in this work, including 113,400 samples of 30 subjects with 15 postures. Extensive experiments on the public MSRA hand gesture dataset, body pose dataset and the proposed writing posture dataset demonstrate that 3D PostureNet achieves significantly superior performance on both skeleton-based human posture and hand posture recognition tasks. (C) 2020 Elsevier B.V. All rights reserved.
资助项目	Major Project for New Generation of AI[2018AAA0100400] ; National Key Research and Development Program[2016YFB0501100] ; National Natural Science Foundation of China[91646207] ; National Natural Science Foundation of China[61976208]
WOS关键词	SYSTEM
WOS研究方向	Computer Science
语种	英语
出版者	ELSEVIER
WOS记录号	WOS:000595366500001
资助机构	Major Project for New Generation of AI ; National Key Research and Development Program ; National Natural Science Foundation of China
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/42704]
专题	自动化研究所_模式识别国家重点实验室_遥感图像处理团队
通讯作者	Wang, Ying
作者单位	1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China
推荐引用方式 GB/T 7714	Liu, Jianbo,Wang, Ying,Liu, Yongcheng,et al. 3D PostureNet: A unified framework for skeleton-based posture recognition[J]. PATTERN RECOGNITION LETTERS,2020,140:143-149.
APA	Liu, Jianbo,Wang, Ying,Liu, Yongcheng,Xiang, Shiming,&Pan, Chunhong.(2020).3D PostureNet: A unified framework for skeleton-based posture recognition.PATTERN RECOGNITION LETTERS,140,143-149.
MLA	Liu, Jianbo,et al."3D PostureNet: A unified framework for skeleton-based posture recognition".PATTERN RECOGNITION LETTERS 140(2020):143-149.