Motor Imagery Signal Classification Based on HDP AR HMM
1 School of Automation Science and Engineering, South ChinaUniversity of Technology, Guangzhou 510640, China
2 Huangshi Electric Power Supply Company of State Grid Corporation of China, Huangshi 435000, China
Abstract Hidden Markov model (HMM) is well applied in brain computer interface, especially in the classification of motor imagery(MI) electroencephalogram (EEG) signal. Conventional methods use HMM to model EEG signal, then use the observed signal under controlled state to estimate the HMM parameters and finally classify the EEG signal through the trained HMMs. However, due to the characteristics of low signaltonoise ratio(SNR), high dimensionality and complexity of motor imagery EEG signal, HMM cannot fully describe the dynamic property of motor imagery EEG signal. In this paper, we use hierarchical Dirichlet process (HDP) with selfclustering ability to describe MI signals and then use AR/VAR model to highlight the time property of MI signal. Finally we combine them with Markov switching processing (MSP) so that we can get more information of MI signal. In order to verify this method, we tested the algorithm on our inhouse data and some of the 2003 BCI international competition data sets. High accuracy on classification of MI is obtained.
WU Jun1 YANG Ya2 YU ZhuLiang1#* GU Zheng Hui1 LI YuanQing1#. Motor Imagery Signal Classification Based on HDP AR HMM[J]. journal1, 2014, 33(6): 666-672.
［1］Vidal JJ. Toward direct braincomputer communication ［J］. Annual Review of Biophysics and Bioengineering, 1973, 2(1):157-180.
［2］Wolpaw JR, McFarland DJ, Vaughan TM. Braincomputer interface research at the Wadsworth Center ［J］. IEEE Transactions on Rehabilitation Engineering, 2000,8(2):222-226.
［3］高上凯．浅谈脑－机接口的发展现状与挑战［J］.中国生物医学工程学报，2007，26(6): 801-803.
［4］尧德中，刘铁军，雷旭，等．基于脑电的脑机接口: 关键技术和应用前景［J］.电子科技大学学报，2009，38(5): 550-554．
［5］Obermaier B, Neuper C, Guger C, et al. Information transfer rate in a fiveclasses braincomputer interface ［J］. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2001, 9(3):283-288.
［6］Bishop CM, Nasser MN. Pattern Recognition and Machine Learning ［M］. New York: Springer, 2006.
［7］Obermaier BG, Guger C, Neuper C, et al. Hidden Markov models for online classification of single trial EEG data ［J］. Pattern Recognition Letters,2001,22(12): 1299-1309. ［8］Rabiner L. A tutorial on hidden Markov models and selected applications in speech recognition ［J］. Proceedings of the IEEE, 1989, 77(2):257-286.
［9］Faisan S, Thoraval L, Armspach JP, et al. Unsupervised learning and mapping of active brain functional MRI signals based on hidden semiMarkov event sequence models ［J］. IEEE Transactions on Medical Imaging, 2005, 24(2):263-276.
［10］Fox EB, Sudderth EB, Jordan MI, et al. A sticky HDPHMM with application to speaker diarization ［J］.The Annals of Applied Statistics, 2011, 5(2A):1020-1056.
［11］Fox EB, Sudderth EB, Jordan MI, et al. Bayesian nonparametric methods for learning Markov switching processes ［J］. IEEE Signal Processing Magazine, 2010,27(6): 43-54.
［12］Willsky AS, Sudderth EB, Jordan MI, et al. Nonparametric Bayesian learning of switching linear dynamical systems ［J］.Advances in Neural Information Processing Systems, 2008,21:457-464.
［13］Yang Y, Yu Z, Gu Z, et al. A new method for motor imagery classification based on hidden Markov model ［C］//Industrial Electronics and Applications (ICIEA). Singapore city: IEEE, 2012:1588-1591.
［14］Lu R, David BD, Lawrence C. The dynamic hierarchical Dirichlet process［C］//Proceedings of the 25th International Conference on Machine Learning. Helsinki: ACM, 2008:824-831.
［15］周建英, 王飞跃, 曾大军. 分层 Dirichlet 过程及其应用综述［J］. 自动化学报, 2011,37(4):389-407.
［16］Mousavi SR, Niknazar MB. Vosoughi V. Epileptic seizure detection using AR model on EEG signals［C］ //Biomedical Engineering Conference. Cairo: IEEE, 2008:1-4.
［17］Neal RM. Markov chain sampling methods for Dirichlet process mixture models ［J］.Journal of Computational and Graphical Statistics, 2000,9(2): 249-265. ［18］Xu T, Zhang Z, Long B, et al. Evolutionary clustering by hierarchical dirichlet process with hidden markov state［C］ //International Conference on Data Mining.Pisa: IEEE,2008:658-667.
