Exercise Recognition Using Averaged Hidden Markov Models
This paper presents a novel learning algorithm for Hidden Markov Models (HMMs) based on multiple learning sequences. For each activity a few left-to-right HMMs are created and then averaged into singular model. Averaged models’ structure is defined by a proposed Sequences Concatenation Algorithm which has been included in this paper. Also the modification of action recognition algorithm for such averaged models has been described.
The experiments have been conducted for the problem of modeling and recognition of chosen 13 warm-up exercises. The input data have been collected using the depth sensor Microsoft Kinect 2.0. The experiments results confirm that an averaged model combines the features of all component models and thus recognizes more sequences. The obtained models do not confuse modeled activities with others.
KeywordsAction recognition Left-to-right Hidden Markov Models Kinect Averaged models
This work was supported by the statutory funds of the Faculty of Electronics 0402/0104/16, Wroclaw University of Science and Technology, Wroclaw, Poland.
- 1.Postawka, A., Śliwiński, P.: A kinect-based support system for children with autism spectrum disorder. In: Rutkowski, L., Korytkowski, M., Scherer, R., Tadeusiewicz, R., Zadeh, L.A., Zurada, J.M. (eds.) ICAISC 2016. LNCS, vol. 9693, pp. 189–199. Springer, Cham (2016). doi: 10.1007/978-3-319-39384-1_17 Google Scholar
- 3.Pal, M., Saha, S., Konar, A.: Distance matching based gesture recognition for healthcare using microsofts kinect sensor. In: International Conference on Microelectronics, Computing and Communications, pp. 1–6 (2016)Google Scholar
- 4.Postawka, A., Śliwiński, P.: Recognition and modeling of atypical children behavior. In: Rutkowski, L., Korytkowski, M., Scherer, R., Tadeusiewicz, R., Zadeh, L.A., Zurada, J.M. (eds.) ICAISC 2015. LNCS, vol. 9119, pp. 757–767. Springer, Cham (2015). doi: 10.1007/978-3-319-19324-3_68 CrossRefGoogle Scholar
- 5.Hai, P.T., Kha, H.H.: An efficient star skeleton extraction for human action recognition using hidden Markov models. In: IEEE Sixth International Conference on Communications and Electronics, pp. 351–356 (2016)Google Scholar
- 6.Siddiqui, S.A., Snober, Y., Raza, S., Khan, F.M., Syed, T.Q.: Arm gesture recognition on microsoft Kinect Using a Hidden Markov Model-based representations of poses. In: International Conference on Information and Communication Technologies, pp. 1–6 (2015)Google Scholar
- 7.Petkovic, M., Jonker, W., Zivkovic, Z.: Recognizing strokes in tennis videos using hidden Markov models. In: IASTED International Conference on Visualization, Imaging and Image Processing, pp. 512–516 (2001)Google Scholar
- 8.Anbarsanti, N., Prihatmanto, A.S.: Dance modelling, learning and recognition system of aceh traditional dance based on hidden Markov model. In: International Conference on Information Technology Systems and Innovation, pp. 86–92 (2014)Google Scholar
- 9.Feng-Shun Lin, J., Kulić, D.: Segmenting human motion for automated rehabilitation exercise analysis. In: Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp. 2881–2884 (2012)Google Scholar
- 13.Jiang, M., Chen, Y., Zhao, Y., Cai, A.: A real-time fall detection system based on HMM and RVM. In: Proceedings of Visual Communications and Image Processing, pp. 1–6 (2013)Google Scholar
- 14.Davis, R.I.A., Walder, C.J., Lovell, B.C.: Improved classification using hidden Markov averaging from multiple observation sequences. In: Proceedings of the Fourth Australasian Workshop on Signal Processing and Applications, pp. 89–92 (2002)Google Scholar