1. Shandong Provincial Key Laboratory of Network Based Intelligent Computing, School of Information Science and Engineering, University of Jinan, Ji'nan 250022, China;
2. School of Computer Science and Technology, Shandong University, Ji'nan 250061, China;
3. School of Control Science and Engineering, Shandong University, Ji'nan 250061, China
To obtain high recognition accuracy of human actions in daily life under different illumination conditions and angles of view, a recognition method based on Kinect is proposed. Inspired by the fact that the human's attention focuses on the detailed partial action in most cases of action recognitions, the proposed recognition method processes the sensed joints data hierarchically. Roughly, actions are classified into the upper limb action and the trunk action by judging the change speed of the position of the trunk joint. Particularly, the hand joints trajectories in the upper limb action and the trunk joint trajectory in the trunk action are focused on. Then, key points are extracted from the two rough categories of trajectories by C-means clustering algorithm for each category separately. And the trajectory of the action is mapped to the corresponding key points. By this means, the sequence of key points in each category is obtained. And the sequences of key points are modeled based on the concept of the proposed temporal order histogram. Nextly, the action recognition is accomplished by comparing the similarities of the sequences of the key points among the trajectories. Finally, the recognition rate of 99% is realized on the collected datasets. The results show that the method is effective for human action recognition task.
[1] 魏莱.基于关节点的人体动作识别及姿态分析研究[D].北京:北京邮电大学, 2014.Wei L. The study of human action recognition and pose analysis based on joint points[D]. Beijing: Beijing University of Posts and Telecommunications, 2014.[2] 胡琼, 秦磊, 黄庆明.基于视觉的人体动作识别综述[J].计算机学报, 2013, 36(12):2512-2524.Hu J, Qin L, Huang Q M. Overview of human action recognition based on vision[J]. Chinese Journal of Computers, 2013, 36(12): 2512-2524.[3] Wang H, Kläser A, Schmid C. Dense trajectories and motion boundary descriptors for action recognition[J]. International Journal of Computer Vision, 2013, 103(1): 60-79. [4] Hoai M, Zisserman A. Improving human action recognition using score distribution and ranking[M]//Lecture Notes in Computer Science, vol.9003. Berlin, Germany: Springer-Verlag, 2015: 3-20.[5] 顾颖媛.稀疏表示在基于视频的人体动作识别中的应用研究[D].北京:清华大学, 2013.Gu Y Y. Application of sparse representation in human motion recognition based on video[D]. Beijing: Tsinghua University, 2013.[6] Simonyan K, Zisserman A. Two-stream convolutional networks for action recognition in videos[C]//Advances in Neural Information Processing Systems. Montreal, Canada: Morgan Kaufmann, 2014: 568-576.[7] 李红波, 丁林建, 冉光勇.基于 Kinect 深度图像的人体识别分析[J].数字通信, 2012(4):21-26.Li H B, Ding L J, Ran G Y. Analysis of human body recognition based on Kinect depth image[J]. Digital Communication, 2012(4): 21-26.[8] Warade S, Aghav J, Claude P, et al. Real-time detection and tracking with Kinect[C]//12th IEEE International Conference on Computer and Information Technology. Piscataway, USA: IEEE, 2012: 86-89.[9] 尹建芹, 田国会, 周风余.智能空间下基于时序直方图的人体行为表示与理解[J].计算机学报, 2014, 37(2):470-479.Yin J Q, Tian G H, Zhou F Y. Human behavior representation and understanding based on temporal histogram in smart space[J]. Chinese Journal of Computers, 2014, 37(2): 470-479.[10] 田国会, 尹建芹, 韩旭, 等.一种基于关节点信息的人体行为识别新方法[J].机器人, 2014, 36(3):285-292.Tian G H, Yin J Q, Han X, et al. A new method of human behavior recognition based on joint point information[J]. Robot, 2014, 36(3): 285-292.[11] 陈曦,孟庆虎.骨架关节点跟踪的人体行为识别方法[J]. 河南科技大学学报:自然科学版,2015,36(2):43-48. Chen X, Meng Q H. Human behavior recognition method for skeleton joint tracking[J]. Journal of Henan University of Science and Technology: Natural Science Edition, 2015, 36(2): 43-48.[12] 尹建芹,田国会,姜海涛,等.面向家庭服务的人体动作 识别[J].四川大学学报:工程科学版,2011,43(4):101-107. Yin J Q, Tian G H, Jiang H T, et al. Human action recognition for family services[J]. Journal of Sichuan University: Engineering Science Edition, 2011, 43(4): 101-107.[13] Ji X F, Wu Q Q, Ju Z J, et al. Study of human action recognition based on improved spatiotemporal features[J]. International Journal of Automation & Computing, 2014, 15(5): 500-509.[14] Yao B, Hagras H, Alhaddad M J, et al. A fuzzy logic-based system for the automation of human behavior recognition using machine vision in intelligent environments[J]. Soft Computing, 2015(19): 499-506.[15] 朱岩,赵旭,刘允才.基于稀疏编码和局部时空特征的人 体动作识别[C]//第十五届全国图象图形学学术会议论文集. 北京:清华大学出版社,2010:237-241. Zhu Y, Zhao X, Liu Y C. Human action recognition based on sparse encoding and local temporal and spatial features[C]//Proceedings of the 15th National Symposium on Image and Graphics. Beijing: Tsinghua University Press, 2010: 237-241.[16] 姬晓飞,王策,李一波.基于一种视角鲁棒性特征的人 体动作识别方法研究[C]//中国自动化学会控制理论专业委 员会、中国系统工程学会第三十二届中国控制会议论文 集(C卷).北京:北京航空航天大学出版社,2013:3877-3881. Ji X F, Wang C, Li Y B. Research on human action recognition method based on the robust feature of a visual angle[C]// Chinese Society of Automation Control Theory, the Chinese System Engineering Society. The 32nd Chinese Control Conference: Vol. C. Beijing: Beihang University Press, 2013: 3877-3881.[17] 叶喜勇,陶霖密,王国建,等.视角无关的人体躯干动 作识别[C]//第六届和谐人机环境联合学术会议(HHME 2010)、第19届全国多媒体学术会议(NCMT2010)、第6届 全国人机交互学术会议(CHCI2010)、第5届全国普适计算 学术会议(PCC2010)论文集.2010. Ye X Y, Tao L M, Wang G J, et al. A novel view invariant algorithm for human main body action recognition[C]//HHME2010, NCMT2010; CHCI2010; PCC2010 Vol. 2010.