基于自学习特征与HMM的人体动作识别

系统仿真学报 ›› 2015, Vol. 27 ›› Issue (8): 1782-1789.

基于自学习特征与HMM的人体动作识别

温加睿, 刘丽娜, 芮玲, 马世伟

上海大学机电工程与自动化学院上海市电站自动化重点实验室,上海 200072

收稿日期:2015-06-15 修回日期:2015-06-24 出版日期:2015-08-08 发布日期:2020-08-03

Human Action Recognition Based on Self-Learning Feature and HMM

Wen Jiarui, Liu Lina, Rui Ling, Ma Shiwei

Shanghai University, School of Mechatronic Engineering and Automation,Shanghai Key Laboratory of Power Station Automation, Shanghai 200072, China

Received:2015-06-15 Revised:2015-06-24 Online:2015-08-08 Published:2020-08-03
About author:Wen Jiarui (1989-), Male, Guangdong, China, Master, Research Filed: Pattern Recognition, Image Processing, Behavior Recognition;Liu Lina (1981-)，Female, Shandong, China, Doctor, University Lecturer, Research Filed: Pattern Recognition, Image Processing, Face Recognition

摘要/Abstract

摘要： 利用机器视觉进行人体动作识别的方法大多数基于手工特征并需要先验知识,这类方法不可避免地依赖于特定问题而忽略了视觉信息的内在结构。提出了一种利用自学习特征及姿态组合规则进行有效动作识别的新方法。使用稀疏自编码(SAE)网络提取轮廓图像的结构特征并构造姿态码本。在识别阶段,使用隐马可夫模型(HMM)训练不同动作类别的模型。设计了一种关键帧提取算法用于在训练HMM前降低长序列的冗余度。通过仿真实验验证了该方法的有效性。

关键词: 动作识别, 自学习特征, SAE, HMM, 姿态码本

Abstract: The current methods of human action recognition by computer vision are mostly based on hand-craft features and usually prior knowledge-required. They inevitably depend on specific applications and neglect the inner structure of visional information. A novel method which integrated self-learned pose features and combined posture symbol rules was proposed to achieve the recognition of human action more efficiently. The structural features of posture silhouette were extracted and a codebook of primary posture was built through the establishment of a sparse auto-encoder network. Then, in the phase of recognition, the Hidden Markov Model was employed to train the models for different action categories. Besides, a key frame extraction algorithm was developed to reduce the redundancy of long code sequence before training HMMs. Simulation experiments manifest the effectiveness of the proposed method.

Key words: pose recognition, self-learned feature, SAE, HMM, posture codebook

中图分类号:

TP391.1

温加睿, 刘丽娜, 芮玲, 马世伟. 基于自学习特征与HMM的人体动作识别[J]. 系统仿真学报, 2015, 27(8): 1782-1789.

Wen Jiarui, Liu Lina, Rui Ling, Ma Shiwei. Human Action Recognition Based on Self-Learning Feature and HMM[J]. Journal of System Simulation, 2015, 27(8): 1782-1789.

参考文献

[1] GU Jun2xia, DING Xiao2qing, WANG Sheng2jin A Survey of Activity Analysis Algorithms[J]. Journal of Image and Graphics(S2301-3699), 2009, 14(3): 377-387.
[2] Liang Penghua.Research of Human Action Recognition Based on HMM[D]. China: Lanzhou University, 2012.
[3] YU Hui, GUO Li.A Posture Description based on Star-skeleton and HMM[J]. Communications Technology (S1002-0802), 2012, 45(12): 91-94.
[4] Li-li Zheng. the Study of Abnormal Human Behavior Detection Algorithm Based on SVM[D]. China: Zhejiang University of Technology, 2012.
[5] Ahad MAR, Ogata T, Tan JK, et al. Moment-based human motion recognition from the representation of DMHI templates[C] //USA: Sice Conference. IEEE, 2008:578-583.
[6] Zhang Xu-dong, et al. Human Activity Recognition Using Multi-layered Motion History Images with Time-Of-Fligh (TOF) Camera[J]. Journal of Electronics and Information Technology(S1064-2269, 2014, 36(5): 1139-1144.
[7] Tian J, Li L, Liu W.Multi-Scale Human Pose Tracking in 2D Monocular Images[J]. Journal of Computer & Communications(S2027-5227), 2014, 2(2): 78-84.
[8] Li Ning, XuDe, Fu Xiaoying, et al. Action Recognition Combined with Human Action Property[J]. Journal of Beijing Jiaotong University(S1673-2-0291), 2009, 33(2): 6-10.
[9] Andrew Ng.Sparse autoencoder. CS294A Lecture notes, 2011: 72
[10] Bengio Y, Lamblin P.Greedy layer-wise training of deep networks[C] //Nips. 2007.
[11] Y Lecun, L Bottou, Y Bengio, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE(S0018-2919), 1998, 86(11): 2278-2324.
[12] J, Xiang L, Hang R, et al. Stacked Sparse Autoencoder (SSAE) based framework for nuclei patch classification on breast cancer histopathology[C] //China. Biomedical Imaging (ISBI), 2014 IEEE 11th International Symposium on. IEEE, 2014:999-1002.
[13] Xu Suping.Research on Human Detection based on Feature Learning in Depth Image[D]. China: Xiamen University, 2014.
[14] Rabiner L R.A tutorial on hidden Markov models and selected applications in speech recognition[J]. Proceedings of the IEEE(S0018-2919), 1989, 77(2): 257-286.
[15] YC Lin, MC Hu, WH Cheng, et al. Human Action Recognition and Retrieval Using Sole Depth Information[DB/OL]. (2012-12-02), [2015-06-15], http://mclab.citi.sinica.edu.tw/dataset/dha/dha.html#classification_database. Proc. 20th ACM Int’l Conf.Multimedia (MM'12), 2012, Nara, Japan.