Myoelectric signal segmentation and classification using wavelets based neural networks

In this paper a method for Myoelectric signal (MES) segmentation and classification is proposed. The classical moving average technique augmented with Principal Components Analysis (PCA), and time-frequency analysis were used for segmentation. Multiresolution Wavelet Analysis (MRWA) was adopted as an effective feature extraction technique while Artificial Neural Networks (ANN) was used for MES classification. Results of classifying four elbow and wrist movements gave 94.9% sensitivity and 94.9% positive predictivity.

[1]  I. Johnstone,et al.  Adapting to Unknown Smoothness via Wavelet Shrinkage , 1995 .

[2]  Erkki Oja,et al.  Subspace methods of pattern recognition , 1983 .

[3]  Ali M. S. Zalzala,et al.  Myoelectric signal classification using evolutionary hybrid RBF-MLP networks , 2000, Proceedings of the 2000 Congress on Evolutionary Computation. CEC00 (Cat. No.00TH8512).

[4]  Stéphane Mallat,et al.  Characterization of Signals from Multiscale Edges , 2011, IEEE Trans. Pattern Anal. Mach. Intell..

[5]  K. Englehart,et al.  Classification of the myoelectric signal using time-frequency based representations. , 1999, Medical engineering & physics.

[6]  Kevin Englehart,et al.  A MICROPROCESSOR-BASED MULTIFUNCTION MYOELECTRIC CONTROL SYSTEM , 2000 .

[7]  Eric Majani,et al.  Biorthogonal wavelets for image compression , 1994, Other Conferences.

[8]  R.N. Scott,et al.  Two-channel enhancement of a multifunction control system , 1995, IEEE Transactions on Biomedical Engineering.

[9]  N Kumaravel,et al.  Automatic diagnosis of neuro-muscular diseases using neural network. , 1994, Biomedical sciences instrumentation.

[10]  R N Scott,et al.  Myoelectric prostheses: state of the art. , 1988, Journal of medical engineering & technology.

[11]  K. Kuribayashi,et al.  A discrimination system using neural network for EMG-controlled prostheses , 1992, [1992] Proceedings IEEE International Workshop on Robot and Human Communication.