Abstract：In order to reconstruct 3D human pose from multi-view images accurately and effectively,a novel human pose estimation algorithm based on multi-kernel sparse coding is proposed.First,for the ambiguity of human pose estimation between the consecutive frames,we describe multi-view images using a special HA-SIFT descriptor,in which the human body local topology,relative coordinates and appearance information are encoded simultaneously;then,an objective function is established within the framework of multi-kernel learning,it takes both intrinsic manifold structure of the feature space and geometrical information of the pose space into consideration.The sparse coding,over-complete dictionary and multi-kernel weight are updated by optimizing the objective function iteratively in the Hilbert space;finally,the corresponding 3D human pose of the unknown input image is estimated by a linear combination of the bases of the human pose dictionary.The experimental results show that our proposed method provides higher accuracy of human pose estimation compared with kernel sparse coding,Laplace sparse coding and Bayesian sparse coding.
余家林, 孙季丰, 李万益. 基于多核稀疏编码的三维人体姿态估计[J]. 电子学报, 2016, 44(8): 1899-1908.
YU Jia-lin, SUN Ji-feng, LI Wan-yi. 3D Human Pose Estimation Based on Multi-kernel Sparse Coding. Acta Electronica Sinica, 2016, 44(8): 1899-1908.
[1] Zhu Ren,Yuan Junsong,Meng Jingjing,et al.Robust part-based hand gesture recognition using kinect sensor[J].IEEE Transaction on Multimedia,2013,15(5):1110-1120. [2] Lin Weiyao,Chen Yuanzhe,Wu Jianxin,et al.A new network-based algorithm for human activity recognition in videos[J].IEEE Transaction on Circuit and System for Video Technology,2014,24(5):826-841. [3] S Vantigodi,W B Radhakrishnan.Action recognition from motion capture data using meta-cognitive RBF network classifier[A].Proceedings of IEEE International Conference on Intelligent Sensors,Sensor Networks and Information Processing (ISSNIP)[C].Singapore:IEEE Computer Society,2014.1-6. [4] G Palmas,M Bachynskyi,A Oulasvirta,et al.MovExp:A versatile visualization tool for human-computer interaction studies with 3D performance and biomechanical data[J].IEEE Transaction on Visualization and Computer Grapgics,2014,20(12):2359-2368. [5] A Agarwal,B Triggs.Recovering 3D human pose from monocular images[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2006,28(1):44-58. [6] J Shotton,R Girshick,A Fitzgibbon,et al.Efficient human pose estimation from single depth images[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2013,35(12):2821-2839. [7] N C Tang,Chiou-Ting Hsu,Weng Mingfang,et al.Example-based human motion extrapolation and motion repairing using contour manifold[J].IEEE Transaction on Multimedia,2014,16(1):47-59. [8] Wang Jing,Lu Canyi,Wang Meng,et al.Robust face recognition via adaptive sparse representation[J].IEEE Transaction on Cybernetics,2014,44(12):2368-2378. [9] 王瑞,杜林峰,孙督等.复杂场景下结合SIFT与核稀疏表示的交通目标分类识别[J].电子学报,2014,42(11):2129-2134.WANG Rui,DU Lin-feng,SUN Du,et al.Traffic object recognition in complex scenes based on SIFT and kernel sparse representation[J].Acta Electronica Sinica,2014,42(11):2129-2134.(in Chinese) [10] Zhou Liuyang,Lu Zhiwu,Howard Leung,et al.Spatial temporal pyramid matching using temporal sparse representation for human motion retrieval[A].Proceedings of International Conference on Computer Graphics International (CGI)[C].Sydney,Australia:Springer,Berlin,2014.845-854. [11] B M Behnam,A Jourabloo,A Zarghami,et al.A Bayseian framework for sparse representation-based 3D human pose estimation[J].IEEE Signal Processing Letters,2014,21(3):297-300. [12] Gao Shenghua,I W Tsang,Chia Liang-Tien.Laplace sparse coding,Hypergraph laplacian sparse coding,and application[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2013,35(1):92-101. [13] Gao Shenghua,I W Tsang,Chia Liang-Tien.Sparse representation with kernels[J].IEEE Transaction on Image Processing,2013,22(2):423-434. [14] Hong Zeng,Yiu-ming Cheung.Feature selection and kernel learning for local learning-based clustering[J].IEEE Transaction on Pattern Analysis and Machine Intelligence,2011,33(8):1532-1546. [15] 赵宏伟,李清亮,刘萍萍.基于分级显著信息的空间编码方法[J].电子学报,2014,42(9):1863-1867.ZHAO Hong-wei,LI Qing-liang,LIU Ping-ping.Spatial encoding based on hierarchical salient information[J].Acta Electronica Sinica,2014,42(9):1863-1867.(in Chinese) [16] Edgar,Simo-Serra,A perez-Gracia.Kinematic synthesis using tree topologies[J].Mechanism and Machine Theory,2014,72(2):94-113. [17] Y Xia,S Changyin.A novel neural dynamical approach to convex quadratic program and its efficient application[J].Neural Network,2009,10(22):1463-1470. [18] Fan Qibin,Jiao Yuling,Lu Xiliang.A primal dual active algorithm with continuation for compressed sensing[J].IEEE Transaction on Signal Processing,2014,62(23):6274-6284. [19] L Sigal,and M J Black.Humaneva:synchronized video and motion capture dataset for evaluation of articulated human motion[R].Report of Brown University,Providence:2006.
2016, Vol. 44 
