Learning a Robust Part-Aware Monocular 3D Human Pose Estimator via Neural Architecture Search

doi:10.1007/s11263-021-01525-0

CASIA OpenIR > 智能感知与计算研究中心

	Learning a Robust Part-Aware Monocular 3D Human Pose Estimator via Neural Architecture Search
	Chen, Zerui1,2 ; Huang, Yan1 ; Yu, Hongyuan1,2 ; Wang, Liang1,2,3,4
发表期刊	INTERNATIONAL JOURNAL OF COMPUTER VISION
ISSN	0920-5691
	2021-10-26
页码	20
通讯作者	Wang, Liang(wangliang@nlpr.ia.ac.cn)
摘要	Even though most existing monocular 3D human pose estimation methods achieve very competitive performance, they are limited in estimating heterogeneous human body parts with the same decoder architecture. In this work, we present an approach to build a part-aware 3D human pose estimator to better deal with these heterogeneous human body parts. Our proposed method consists of two learning stages: (1) searching suitable decoder architectures for specific parts and (2) training the part-aware 3D human pose estimator built with these optimized neural architectures. Consequently, our searched model is very efficient and compact and can automatically select a suitable decoder architecture to estimate each human body part. In comparison with previous state-of-the-art models built with ResNet-50 network, our method can achieve better performance and reduce 64.4% parameters and 8.5% FLOPs (multiply-adds). We validate the robustness and stability of our searched models by conducting extensive and rigorous ablation experiments. Our method can advance state-of-the-art accuracy on both the single-person and multi-person 3D human pose estimation benchmarks with affordable computational cost.
关键词	Monocular 3D human pose estimation Heterogeneous human body parts Neural architecture search
DOI	10.1007/s11263-021-01525-0
收录类别	SCI
语种	英语
资助项目	National Key Research and Development Program of China[2018AAA0100400] ; National Natural Science Foundation of China[61633021] ; National Natural Science Foundation of China[61721004] ; National Natural Science Foundation of China[61806194] ; National Natural Science Foundation of China[U1803261] ; National Natural Science Foundation of China[61976132] ; Beijing Nova Program[Z201100006820079] ; Shandong Provincial Key Research andDevelopment Program[2019JZZY010119] ; Key Research Program of Frontier Sciences CAS[ZDBS-LY-JSC032] ; CAS-AIR
项目资助者	National Key Research and Development Program of China ; National Natural Science Foundation of China ; Beijing Nova Program ; Shandong Provincial Key Research andDevelopment Program ; Key Research Program of Frontier Sciences CAS ; CAS-AIR
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:000711300700001
出版者	SPRINGER
七大方向——子方向分类	目标检测、跟踪与识别
引用统计
文献类型	期刊论文
条目标识符	http://ir.ia.ac.cn/handle/173211/46271
专题	智能感知与计算研究中心
通讯作者	Wang, Liang
作者单位	1.CASIA, Ctr Res Intelligent Percept & Comp, NLPR, Beijing, Peoples R China 2.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing, Peoples R China 3.Chinese Acad Sci, Ctr Excellence Brain Sci & Intelligence Technol, Beijing, Peoples R China 4.Chinese Acad Sci, Artificial Intelligence Res, Beijing, Peoples R China
第一作者单位	模式识别国家重点实验室
通讯作者单位	模式识别国家重点实验室
推荐引用方式 GB/T 7714	Chen, Zerui,Huang, Yan,Yu, Hongyuan,et al. Learning a Robust Part-Aware Monocular 3D Human Pose Estimator via Neural Architecture Search[J]. INTERNATIONAL JOURNAL OF COMPUTER VISION,2021:20.
APA	Chen, Zerui,Huang, Yan,Yu, Hongyuan,&Wang, Liang.(2021).Learning a Robust Part-Aware Monocular 3D Human Pose Estimator via Neural Architecture Search.INTERNATIONAL JOURNAL OF COMPUTER VISION,20.
MLA	Chen, Zerui,et al."Learning a Robust Part-Aware Monocular 3D Human Pose Estimator via Neural Architecture Search".INTERNATIONAL JOURNAL OF COMPUTER VISION (2021):20.