Multi-channel electromyography pattern classification using deep belief networks for enhanced user experience

Abstract

An enhanced algorithm is proposed to recognize multi-channel electromyography (EMG) patterns using deep belief networks (DBNs). It is difficult to classify the EMG features because an EMG signal has nonlinear and time-varying characteristics. Therefore, in several previous studies, various machine-learning methods have been applied. A DBN is a fast, greedy learning algorithm that can find a fairly good set of weights rapidly, even in deep networks with a large number of parameters and many hidden layers. To evaluate this model, we acquired EMG signals, extracted their features, and then compared the model with the DBN and other conventional classifiers. The accuracy of the DBN is higher than that of the other algorithms. The classification performance of the DBN model designed is approximately 88.60%. It is 7.55% (p=9.82×10⁻¹²) higher than linear discriminant analysis (LDA) and 2.89% (p=1.94×10⁻⁵) higher than support vector machine (SVM). Further, the DBN is better than shallow learning algorithms or back propagation (BP), and this model is effective for an EMG-based user-interfaced system.

Keywords

electromyography (EMG) / pattern classification / feature extraction / deep learning / deep belief network (DBN)

References

[1] HudginsB, ParkerP, ScottR. A new strategy for multifunction myoelectric control [J]. IEEE Trans on Biomedical Engineering, 1993, 40(1): 82-94

[2] LeeJ, LeeG. Gait angle prediction for lower limb orthotics and prostheses using an EMG signal and neural networks [J]. International Journal of Control, Automation, and Systems, 2005, 3(2): 152-158

[3] AjiboyeA, WeirR. A heuristic fuzzy logic approach to EMG pattern recognition for multifunctional prosthesis control [J]. IEEE Trans on Neural Systems and Rehabilitation Engineering, 2005, 13(3): 280-291

[4] KhokharZ O, XiaoZ G, MenonC. Surface EMG pattern recognition for real-time control of a wrist exoskeleton [J]. BioMedical Engineering OnLine, 2010, 9: 41

[5] ChenL, GengY, LiG. Effect of upper-limb positions on motion pattern recognition using electromyography [C]. 4th International Congress on Image and Signal Processing (CISP), 2011, Shanghai, IEEE Press: 139-142

[6] SchemeE, EnglehartK. Electromyogram pattern recognition for control of powered upper-limb prostheses: State of the art and challenges for clinical use [J]. Journal of Rehabilitation Research & Development, 2011, 48(6): 643-660

[7] LorrainT, JiangN, FarinaD. Influence of the training set on the accuracy of surface EMG classification in dynamic contractions for the control of multifunction prostheses [J]. Journal of NeuroEngineering and Rehabilitation, 2011, 8(1): 25

[8] YoungA J, SmithL H, RouseE J, HargroveL J. Classification of simultaneous movements using surface EMG pattern recognition [J]. IEEE Trans on Biomedical Engineering, 2013, 60(5): 1250-1258

[9] NingJ, RehbaumH, VujaklijaI, GraimannB, FarinaD. Intuitive, online, simultaneous, and proportional myoelectric control over two degrees-of-freedom in upper limb amputees [J]. IEEE Trans on Neural Systems and Rehabilitation Engineering, 2014, 22(3): 501-510

[10] MandrykR L, InkpenK M, CalvertT W. Using psychophysiological techniques to measure user experience with entertainment technologies [J]. Behaviour & Information Technology, 2006, 25(2): 141-158

[11] AhsanM R, IbrahimyM I, KhalifaO O. EMG signal classification for human computer interaction: A review [J]. European J Scientific Research, 2009, 33(3): 480-501

[12] JaimeG G, IsraelS J G, LuisF N A, SergioA G. Steering a tractor by means of an emg-based human-machine interface [J]. Sensors, 2011, 11(7): 7110-7126

[13] CortesC, VapnikV. Support-vector networks [J]. Machine Learning, 1995, 20(3): 273-297

[14] SubasiA. Classification of EMG signals using PSO optimized SVM for diagnosis of neuromuscular disorders [J]. Computers in Biology and Medicine, 2013, 43: 576-586

[15] ZhangD, XiongA, ZhaoX, HanJ. PCA and LDA for EMG-based control of bionic mechanical hand [C]. International Conference on Information and Automation (ICIA), 2012, Shenyang, IEEE Press: 960-965

[16] KimK S, ChoiH H, MoonC S, MunC W. Comparison of k-nearest neighbor, quadratic discriminant and linear discriminant analysis in classification of electromyogram signals based on the wrist-motion directions [J]. Current Applied Physics, 2011, 11(3): 740-745

[17] HintonG, OsinderoS, TheY. A fast learning algorithm for deep belief nets [J]. Neural Computation, 2006, 18(7): 1527-1554

[18] MohamedA, SainathT N, DahlG, RamabhadranB. Deep belief networks using discriminative features for phone recognition [C]. 2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2011, Prague, IEEE Press: 5060-5063

[19] MohamedA, DahlG E, HintonG. Acoustic modeling using deep belief networks [J]. IEEE Trans on Audio, Speech, and Language Processing, 2012, 20(1): 14-22

[20] ParkS, LeeS. EMG pattern recognition based on artificial intelligence techniques [J]. IEEE Trans on Rehabilitation Engineering, 1998, 6(4): 400-405

[21] JeongE, KimS, SongY, LeeS. Comparison of wrist motion classification methods using surface electromyogram [J]. Journal of Central South University, 2013, 20(4): 960-968

[22] LeeS, KimJ, ParkS. An enhanced feature extraction algorithm for EMG pattern classification [J]. IEEE Trans on Rehabilitation Engineering, 1996, 4(4): 439-443