Karcher均值算法在动作识别上的应用
The Application of Karcher Mean Algorithm in Action Recognition
DOI: 10.12677/PM.2021.116131, PDF,    科研立项经费支持
作者: 赵 倩：上海大学，上海；叶 震, 周朝政：上海电气集团股份有限公司中央研究院，上海
关键词: 动作识别；Karcher均值算法；骨架表示；轨迹时间对齐；Action Recognition； Karcher Mean Algorithm； Skeleton Representation； Trajectory Temporal Misalignment
摘要: 本文提出Karcher均值算法与测地距离相结合的平均轨迹计算方法，并将其应用到动作识别任务中。具体地说，本文通过基于李群的骨架表示方法，将动作骨架序列表示为流形上的轨迹。为了解决轨迹的时间错位问题，本文将Karcher均值算法与李群上测地距离的定义相结合，计算所有动作轨迹的平均轨迹，然后采用传输平方根向量场表示方法，将所有动作轨迹与平均轨迹进行时间对齐。此外，本文在特征提取阶段，提出对特征进行加权融合，实验结果验证了融合特征的有效性。
Abstract: In this paper, we propose an average trajectory calculation method based on the combination of Karcher mean algorithm and geodesic distance, and apply it to the action recognition task. Specif-ically, the skeletal sequences of actions are represented as trajectories on manifolds by a skeleton representation method based on Lie groups. In order to solve the temporal misalignment problem of trajectories, this paper combines the Karcher mean algorithm with the definition of geodesic distance on Lie group, calculates the average trajectories of all action trajectories, and then uses the Transported Square-Root Vector Field representation method to align all action trajectories with the average trajectories in time. In addition, the weighted fusion of features is proposed in the feature extraction stage, and the experimental results verify the effectiveness of the fusion features.
文章引用：赵倩, 叶震, 周朝政. Karcher均值算法在动作识别上的应用[J]. 理论数学, 2021, 11(6): 1166-1180. https://doi.org/10.12677/PM.2021.116131

参考文献

[1] Polat, E., Yeasin, M. and Sharma, R. (2003) Robust Tracking of Human Body Parts for Collaborative Human Computer Interaction. Computer Vision & Image Understanding, 89, 44-69. [Google Scholar] [CrossRef] 
[2] Mokhber, A., Achard, C. and Milgram, M. (2008) Recogni-tion of Human Behavior by Space-Time Silhouette Characterization. Pattern Recognition Letters, 29, 81-89. [Google Scholar] [CrossRef] 
[3] Chang, E. and Wang, Y.F. (2004) Introduction to the Special Issue on Video Surveillance. Multimedia Systems, 10, 116-117. [Google Scholar] [CrossRef] 
[4] 房菲. 基于核方法的支持向量机在人体动作识别中的应用研究[D]: [硕士学位论文]. 青岛: 中国海洋大学, 2013.
[5] Chen, Y., Duff, M., Lehrer, N., et al. (2011) A Computational Framework for Quantitative Evaluation of Movement during Rehabilitation. AIP Conference Proceedings, 1371, 317-326.
[6] Vemulapalli, R., Arrate, F. and Chellappa, R. (2014) Human Action Recognition by Representing 3D Skeletons as Points in a Lie Group. In: Proceed-ings of the IEEE Conference on Computer Vision and Pattern Recognition, IEEE, Columbus, 588-595. [Google Scholar] [CrossRef] 
[7] Su, J., Kurtek, S., Klassen, E., et al. (2014) Statistical Analysis of Trajectories on Riemannian Manifolds: Bird Migration, Hurricane Tracking and Video Surveillance. Annals of Applied Statistics, 8, 530-552. [Google Scholar] [CrossRef] 
[8] Anirudh, R., Turaga, P., Su, J., et al. (2016) Elastic Functional Coding of Riemannian Trajectories. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39, 922-936. [Google Scholar] [CrossRef] 
[9] Amor, B.B., Su, J. and Srivastava, A. (2015) Action Recogni-tion Using Rate-Invariant Analysis of Skeletal Shape Trajectories. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38, 1-13. [Google Scholar] [CrossRef] 
[10] Wang, J., Liu, Z., Wu, Y., et al. (2012) Mining Actionlet En-semble for Action Recognition with Depth Cameras. 2012 IEEE Conference on Computer Vision and Pattern Recogni-tion, Providence, 16-21 June 2012, 1290-1297.
[11] Hall, B. (2015) Lie Groups, Lie Algebras, and Representations: An Elementary Introduction. Springer, Berlin. [Google Scholar] [CrossRef] 
[12] Murray, R.M., Li, Z., Sastry, S.S., et al. (1994) A Mathematical Introduction to Robotic Manipulation. CRC Press, Boca Raton.
[13] Zefran, M. and Kumar, V. (1998) Two Methods for Interpolating Rigid Body Motions. Proceedings 1998 IEEE International Conference on Robotics and Automation, Vol. 4, 2922-2927.
[14] Zefran, M., Kumar, V. and Croke, C. (1996) Choice of Riemannian Metrics for Rigid Body Kinematics. International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, V02BT02A030.
[15] Grove, K. and Karcher, H. (1973) How to Conjugate C1-Close Group Actions. Mathematische Zeitschrift, 132, 11-20. [Google Scholar] [CrossRef] 
[16] Xia, L., Chen, C.C. and Aggarwal, J.K. (2012) View Invariant Human Action Recognition Using Histograms of 3D Joints. 2012 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, Providence, 16-21 June 2012, 20-27. [Google Scholar] [CrossRef] 
[17] Li, W., Zhang, Z. and Liu, Z. (2010) Action Recognition Based on a Bag of 3d Points. 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Francisco, 13-18 June 2010, 9-14. [Google Scholar] [CrossRef] 
[18] Zhu, Y., Chen, W. and Guo, G. (2013) Fusing Spatiotemporal Features and Joints for 3d Action Recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, Portland, 23-28 June 2013, 486-491. [Google Scholar] [CrossRef] 
[19] Zhang, S., Liu, X. and Xiao, J. (2017) On Geometric Features for Skeleton-Based Action Recognition Using Multilayer LSTM Networks. 2017 IEEE Winter Conference on Applications of Computer Vision, Santa Rosa, 24-31 March 2017, 148-157. [Google Scholar] [CrossRef] 
[20] Tanfous, A.B., Drira, H. and Amor, B.B. (2019) Sparse Coding of Shape Trajectories for Facial Expression and Action Recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42, 2594-2607. [Google Scholar] [CrossRef] 

为你推荐