本文提出了一种基于3D人体骨架的动作识别方法.该方法以3D人体骨架为基础，将骨架中关节点的位置重新定义，形成简化的立体骨架模型，进而采用改进的动态时间规整算法（Reformative Dynamic Time Warping，R-DTW）对齐动作序列并进行识别.由于人体大小、形状、动作方式等差异，任意两个人表达同一动作都不尽相同，简化的立体骨架模型能有效缓解这种类内差异性.传统的DTW算法存在计算复杂性高，效率低的问题，本文在传统算法的基础上设计了"一次规划，二次细化"的方法，有效降低计算量，提高计算效率.该算法在MSR 3D Action数据库上的实验验证了其有效性.
This paper presents an action recognition method based on 3D skeleton.This method redefines the coordinates of the articulations which belong to the skeleton to form a simplistic skeleton model firstly.Then a reformative dynamic time warping (R-DTW) algorithm is applied to implement action recognition.There are no two persons identical in an action owing to the difference of body size,shape and action expression.The simplistic skeleton model could decrease this intra-class variability effectively.The drawbacks of conventional DTW algorithm lie in high computational complexity and low recognition efficiency.To solve this problem,we design a method named "Planning & Refining".We conduct this algorithm on MSR Action3D dataset and the results demonstrate its effectiveness.
[1] Wang J,Liu Z,Wu Y.Mining actionlet ensemble for action recognition with depth cameras[A].IEEE Conference on Computer Vision and Pattern Recognition[C].Providence,Rhode Island,USA:IEEE,2012.1290-1297.
[2] Yu G,Liu Z,Yuan J.Discriminative orderlet mining for real-time recognition of human-object interaction[A].Asian Conference on Computer Vision[C].Singapore:Springer,2014.50-65.
[3] Li W,Zhang Z,Liu Z.Action recognition based on a bag of 3d points[A].IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops[C].San Francisco,California,USA:IEEE,2010.9-14.
[4] Xia L,Chen C C,Aggarwal J K.View invariant human action recognition using histograms of 3D joints[A].IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops[C].Providence,Rhode Island,USA:IEEE,2012.20-27.
[5] Johansson G.Visual motion perception[J].Scientific American,1975:76-88.
[6] Ziaeefard M,Ebrahimnezhad H.Hierarchical human action recognition by normalized-polar histogram[A].International Conference on Pattern Recognition[C].Istanbul,Turkey:IEEE 2010.3720-3723.
[7] Itakura F.Minimum prediction residual principle applied to speech recognition[J].IEEE Transactions on Acoustics,Speech and Signal Processing,1975,23(1):67-72.
[8] Morales-Cordovilla J A,Cabanas-Molero P,Peinado A M,et al.A robust pitch extractor based on DTW lines and CASA with application in noisy speech recognition[A].Advances in Speech and Language Technologies for Iberian Languages[M].Heidelberg,Berlin:Springer,2012.197-206.
[9] Ikizler N,Duygulu P.Histogram of oriented rectangles:A new pose descriptor for human action recognition[J].Image and Vision Computing,2009,27(10):1515-1526.
[10] Candan K S,Rossini R,Wang X,et al.sDTW:computing DTW distances using locally relevant constraints based on salient feature alignments[J].Proceedings of the VLDB Endowment,2012,5(11):1519-1530.
[11] Zhang W,Zhang Y,Gao C,et al.Action recognition by joint spatial-temporal motion feature[J].Journal of Applied Mathematics,2013,2013(1):467-471.
[12] Ruan X,Tian C.Dynamic gesture recognition based on improved DTW algorithm[A].IEEE International Conference on Mechatronics and Automation[C].Beijing,China:IEEE,2015.2134-2138.
[13] Theodorakopoulos I,Kastaniotis D,Economou G,et al.Pose-based human action recognition via sparse representation in dissimilarity space[J].Journal of Visual Communication and Image Representation,2014,25(1):12-23.
[14] 徐贵力,赵妍,姜斌,等.基于局部尺度特征描述和改进DTW技术的局部轮廓匹配算法[J].电子学报,2016,44(1):135-142. Xu GuiLi,Zhao Yan,Jiang Bin,et al.Partial contour matching algorithm based on local scale description and improved DTW[J].Acta Electronica Sinica,2016,44(1):135-142.(in Chinese)
2017, Vol. 45 
