A speech recognition system based on electromyography for the rehabilitation of dysarthric patients: A Thai syllable study

• Autorzy: Sae Jong Nida , Phukpattaranont Pornchai

Identyfikatory

DOI

10.1016/j.bbe.2018.11.010

• Języki publikacji: EN

Abstrakty

EN

The objective of this study is to develop a speech recognition system for classifying nine Thai syllables, which is used for the rehabilitation of dysarthric patients, based on five channels of surface electromyography (sEMG) signals from the human articulatory muscles. After the sEMG signal from each channel was collected, it was processed by a band-pass filter from 20–450 Hz for noise removal. Then, six features from three feature categories were determined and analyzed, namely, mean absolute value (MAV) and wavelength (WL) from amplitude based features (ABF), zero crossing (ZC) and mean frequency (MNF) from frequency based features (FBF), and L-kurtosis (L-KURT) and L-skewness (L-SKW) from statistics based features (SBF). Subsequently, a spectral regression extreme learning machine (SRELM) was used as the feature projection technique to reduce the dimension of feature vector from 30 to 8. Finally, the projected features were classified using a feed forward neural network (NN) classifier with 5-fold cross-validation. The proposed system was evaluated with the sEMG signals from seven healthy volunteers and five dysarthric volunteers. The results show that the proposed system can recognize the sEMG signals from both healthy and dysarthric volunteers. The average classification accuracies obtained from all six features in the healthy and dysarthric volunteers were 94.5% and 89.4%, respectively.

Słowa kluczowe

EN

dysarthria; surface electromyography; feature classification; speech recognition system

PL

dyzartria; elektromiografia powierzchniowa; system rozpoznawania mowy

• Wydawca: Nałęcz Institute of Biocybernetics and Biomedical Engineering of the Polish Academy of Sciences ; Elsevier
• Czasopismo: Biocybernetics and Biomedical Engineering
• Rocznik: 2019
• Tom: Vol. 39, no. 1
• Strony: 234--245
• Opis fizyczny: Bibliogr. 27 poz., rys., tab., wykr.

Twórcy

autor

Sae Jong Nida

• Department of Electrical Engineering, Faculty of Engineering, Prince of Songkla University, Hat Yai, Songkhla, Thailand

autor

Phukpattaranont Pornchai

• Department of Electrical Engineering, Faculty of Engineering, Prince of Songkla University, Hat Yai, Songkhla, Thailand

Bibliografia

• [1] Mustafa MB, Salim SS, Mohamed N, Al-Qatab B, Siong CE. Severity-based adaptation with limited data for ASR to aid dysarthric speakers. PLoS One 2014;9(5):1–11.
• [2] Kayasith P, Theeramunkong T. Speech confusion index (F): a confusion-based speech quality indicator and recognition rate prediction for dysarthria. Comput Math Appl 2009;58:1534–49.
• [3] Lansford KL, Liss JM. Vowel acoustics in dysarthria: speech disorder diagnosis and classification. J Speech Lang Hear Res 2014;57(1):57–67.
• [4] Kim MJ, Kim Y, Kim H. Automatic intelligibility assessment of dysarthric speech using phonologically-structured sparse linear model. IEEE/ACM Trans Audio Speech Lang Process 2015;23(4):694–704.
• [5] Kim YJ, Weismer G, Kent RD, Duffy JR. Statistical models of F2 slope in relation to severity of dysarthria. Folia Phoniatr Logo 2009;61(6):329–35.
• [6] Kim YJ, Kent RD, Weismer G. An acoustic study of the relationships among neurologic disease, dysarthria type and severity of dysarthria. J Speech Lang Hear Res 2011;54 (2):417–29.
• [7] Darley F, Aronson A, Brown J. Differential diagnostic patterns of dysarthria. J Speech Lang Hear Res 1969;12 (2):246–69.
• [8] Srisuwan N, Phukpattaranont P, Limsakul C. Comparison of feature evaluation criteria for speech recognition based on electromyography. Med Biol Eng Comput 2018;56(6):1041–51.
• [9] Janke M, Diener L. EMG-to-Speech: direct generation of speech from facial electromyographic signals. IEEE/ACM Trans Audio Speech Lang Process 2017;25(12):2375–85.
• [10] Schultz T, Wand M, Hueber T, Krusienski DJ, Herff C, Brumberg JS. Biosignal-based spoken communication: a survey. IEEE/ACM Trans Audio Speech Lang Process 2017;25 (12):2257–71.
• [11] Chan ADC, Englehart KB, Hudgins B, Lovely DF. Myoelectric signals to augment speech recognition. Med Biol Eng Comput 2001;39:500–4.
• [12] Lee KS. EMG-based speech recognition using hidden Markov models with global control variables. IEEE Trans Biomed Eng 2008;55(3):930–40.
• [13] Tsuji T, Bu N, Arita J, Ohga M. A speech synthesizer using facial EMG signals. Int J Comput Intell Appl 2008;7(1):1–15.
• [14] Kubo T, Yoshida M, Hattori T, Ikeda K. Towards excluding redundancy in electrode grid for automatic speech recognition based on surface EMG. Neurocomputing 2014;134:15–9.
• [15] Bunderson NE, Kuiken TA. Quantification of feature space changes with experience during electromyogram pattern recognition control. IEEE Trans Neural Syst Rehabil Eng 2012;20(3):239–46.
• [16] Chan ADC, Englehart K, Hudgins B, Lovely DF. Hidden Markov model classification of myoelectric signals in speech. IEEE Eng Med Biol Mag 2002;21(5):143–6.
• [17] Jou SCS, Schultz T. Automatic speech recognition based on electromyographic biosignals. International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC). 2008. pp. 305–20.
• [18] Anam K, Al-Jumaily A. A novel extreme learning machine for dimensionality reduction on finger movement classification using sEMG. Proceedings of the Seventh International IEEE/EMBS Conference on Neural Engineering (NER); 2015. p. 824–7.
• [19] Phukpattaranont P, Thongpanja S, Anam K, Al-Jumaily A, Limsakul C. Evaluation of feature extraction techniques and classifiers for finger movement recognition using surface electromyography signal. Med Biol Eng Comput 2018;56(12):2259–71.
• [20] Englehart K, Hudgins B. A robust, real-time control scheme for multifunction myoelectric control. IEEE Trans Biomed Eng 2003;50(7):848–54.
• [21] Du YC, Lin CH, Shyu LY, Chen T. Portable hand motion classifier for multi-channel surface electromyography recognition using grey relation analysis. Expert Syst Appl 2010;37(6):4283–91.
• [22] Phinyomark A, Phukpattaranont P, Limsakul C. Feature reduction and selection for EMG signal classification. Expert Syst Appl 2012;39(8):7420–31.
• [23] Du S, Vuskovic M. Temporal vs. spectral approach to feature extraction from prehensile EMG signals. Proceedings of IEEE International Conference on Information Reuse and Integration (IRI); 2004. p. 344–50.
• [24] Thongpanja S, Phinyomark A, Quaine F, Laurillau Y, Limsakul C, Phukpattaranont P. Probability density functions of stationary surface EMG signals in noisy environments. IEEE Trans Instrum Meas 2016;65(7):1547–57.
• [25] Sankarasubramanian A, Srinivasan K. Investigation and comparison of sampling properties of L-moments and conventional moments. J Hydrol 1999;218(1–2):13–34.
• [26] Lapatki BG, Stegeman DF, Jonas IE. A surface EMG electrode for the simultaneous observation of multiple facial muscles. J Neurosci Methods 2003;123(2):117–28.
• [27] Pothirat T. Characterization and analysis of neck and face surface electromyography for speech rehabilitation training.[Master thesis] Prince of Songkla University; 2014.

Uwagi

PL

Opracowanie rekordu w ramach umowy 509/P-DUN/2018 ze środków MNiSW przeznaczonych na działalność upowszechniającą naukę (2019).