Multiday Evaluation of Techniques for EMG Based Classification of Hand Motions

Muhammad Asim Waris, Imran Khan Niazi, Mohsin Jamil, Kevin Englehart, Winnie Jensen, Ernest Nlandu Kamavuako

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

2 Citationer (Scopus)
1 Downloads (Pure)

Resumé

Currently, most of the adopted myoelectric schemes for upper limb prostheses do not provide users with intuitive control. Higher accuracies have been reported using different classification algorithms but investigation on the reliability over time for these methods is very limited. In this study, we compared for the first time the longitudinal performance of selected state-of-the-art techniques for Electromyography (EMG) based classification of hand motions. Experiments were conducted on ten able-bodied and six transradial amputees for seven continuous days. Linear Discriminant Analysis (LDA), Artificial Neural Network (ANN), Support Vector Machine (SVM), K-Nearest Neighbour (KNN) and Decision Trees (TREE) were compared. Comparative analysis showed that the ANN attained highest classification accuracy followed by LDA. Three-way repeated ANOVA test showed a significant difference (P<0.001) between EMG types (surface, intramuscular and combined), Days (1-7), classifiers and their interactions. Performance on last day was significantly better (P<0.05) than the first day for all classifiers and EMG types. Within-day classification error (WCE) across all subject and days in ANN was: surface (9.12 ± 7.38%), intramuscular (11.86±7.84%) and combined (6.11±7.46%). The between-day analysis in a leave-one-day-out fashion showed that ANN was the optimal classifier (surface (21.88 ± 4.14%) intramuscular (29.33 ± 2.58%) and combined (14.37 ± 3.10%)). Results indicate that that within day performances of classifiers may be similar but over time it may lead to a substantially different outcome. Furthermore, training ANN on multiple days might allow capturing time-dependent variability in the EMG signals and thus minimizing the necessity for daily system recalibration. Index Terms— Electromyography; Pattern recognition; Classification; Myoelectric control; Prostheses; Intramuscular
OriginalsprogEngelsk
TidsskriftIEEE Journal of Biomedical and Health Informatics
Vol/bind23
Udgave nummer4
Sider (fra-til)1526-1534
Antal sider11
ISSN1089-7771
DOI
StatusUdgivet - jul. 2019

Fingerprint

artificial neural network
discriminant analysis
pattern recognition
limb
evaluation
experiment
analysis

Citer dette

@article{719a2e42c27a491a8a3981e49546c7d2,
title = "Multiday Evaluation of Techniques for EMG Based Classification of Hand Motions",
abstract = "Currently, most of the adopted myoelectric schemes for upper limb prostheses do not provide users with intuitive control. Higher accuracies have been reported using different classification algorithms but investigation on the reliability over time for these methods is very limited. In this study, we compared for the first time the longitudinal performance of selected state-of-the-art techniques for Electromyography (EMG) based classification of hand motions. Experiments were conducted on ten able-bodied and six transradial amputees for seven continuous days. Linear Discriminant Analysis (LDA), Artificial Neural Network (ANN), Support Vector Machine (SVM), K-Nearest Neighbour (KNN) and Decision Trees (TREE) were compared. Comparative analysis showed that the ANN attained highest classification accuracy followed by LDA. Three-way repeated ANOVA test showed a significant difference (P<0.001) between EMG types (surface, intramuscular and combined), Days (1-7), classifiers and their interactions. Performance on last day was significantly better (P<0.05) than the first day for all classifiers and EMG types. Within-day classification error (WCE) across all subject and days in ANN was: surface (9.12 ± 7.38{\%}), intramuscular (11.86±7.84{\%}) and combined (6.11±7.46{\%}). The between-day analysis in a leave-one-day-out fashion showed that ANN was the optimal classifier (surface (21.88 ± 4.14{\%}) intramuscular (29.33 ± 2.58{\%}) and combined (14.37 ± 3.10{\%})). Results indicate that that within day performances of classifiers may be similar but over time it may lead to a substantially different outcome. Furthermore, training ANN on multiple days might allow capturing time-dependent variability in the EMG signals and thus minimizing the necessity for daily system recalibration. Index Terms— Electromyography; Pattern recognition; Classification; Myoelectric control; Prostheses; Intramuscular",
author = "Waris, {Muhammad Asim} and Niazi, {Imran Khan} and Mohsin Jamil and Kevin Englehart and Winnie Jensen and Kamavuako, {Ernest Nlandu}",
year = "2019",
month = "7",
doi = "10.1109/JBHI.2018.2864335",
language = "English",
volume = "23",
pages = "1526--1534",
journal = "I E E E Transactions on Information Technology in Biomedicine",
issn = "1089-7771",
publisher = "IEEE",
number = "4",

}

Multiday Evaluation of Techniques for EMG Based Classification of Hand Motions. / Waris, Muhammad Asim; Niazi, Imran Khan; Jamil, Mohsin; Englehart, Kevin; Jensen, Winnie; Kamavuako, Ernest Nlandu.

I: IEEE Journal of Biomedical and Health Informatics, Bind 23, Nr. 4, 07.2019, s. 1526-1534.

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

TY - JOUR

T1 - Multiday Evaluation of Techniques for EMG Based Classification of Hand Motions

AU - Waris, Muhammad Asim

AU - Niazi, Imran Khan

AU - Jamil, Mohsin

AU - Englehart, Kevin

AU - Jensen, Winnie

AU - Kamavuako, Ernest Nlandu

PY - 2019/7

Y1 - 2019/7

N2 - Currently, most of the adopted myoelectric schemes for upper limb prostheses do not provide users with intuitive control. Higher accuracies have been reported using different classification algorithms but investigation on the reliability over time for these methods is very limited. In this study, we compared for the first time the longitudinal performance of selected state-of-the-art techniques for Electromyography (EMG) based classification of hand motions. Experiments were conducted on ten able-bodied and six transradial amputees for seven continuous days. Linear Discriminant Analysis (LDA), Artificial Neural Network (ANN), Support Vector Machine (SVM), K-Nearest Neighbour (KNN) and Decision Trees (TREE) were compared. Comparative analysis showed that the ANN attained highest classification accuracy followed by LDA. Three-way repeated ANOVA test showed a significant difference (P<0.001) between EMG types (surface, intramuscular and combined), Days (1-7), classifiers and their interactions. Performance on last day was significantly better (P<0.05) than the first day for all classifiers and EMG types. Within-day classification error (WCE) across all subject and days in ANN was: surface (9.12 ± 7.38%), intramuscular (11.86±7.84%) and combined (6.11±7.46%). The between-day analysis in a leave-one-day-out fashion showed that ANN was the optimal classifier (surface (21.88 ± 4.14%) intramuscular (29.33 ± 2.58%) and combined (14.37 ± 3.10%)). Results indicate that that within day performances of classifiers may be similar but over time it may lead to a substantially different outcome. Furthermore, training ANN on multiple days might allow capturing time-dependent variability in the EMG signals and thus minimizing the necessity for daily system recalibration. Index Terms— Electromyography; Pattern recognition; Classification; Myoelectric control; Prostheses; Intramuscular

AB - Currently, most of the adopted myoelectric schemes for upper limb prostheses do not provide users with intuitive control. Higher accuracies have been reported using different classification algorithms but investigation on the reliability over time for these methods is very limited. In this study, we compared for the first time the longitudinal performance of selected state-of-the-art techniques for Electromyography (EMG) based classification of hand motions. Experiments were conducted on ten able-bodied and six transradial amputees for seven continuous days. Linear Discriminant Analysis (LDA), Artificial Neural Network (ANN), Support Vector Machine (SVM), K-Nearest Neighbour (KNN) and Decision Trees (TREE) were compared. Comparative analysis showed that the ANN attained highest classification accuracy followed by LDA. Three-way repeated ANOVA test showed a significant difference (P<0.001) between EMG types (surface, intramuscular and combined), Days (1-7), classifiers and their interactions. Performance on last day was significantly better (P<0.05) than the first day for all classifiers and EMG types. Within-day classification error (WCE) across all subject and days in ANN was: surface (9.12 ± 7.38%), intramuscular (11.86±7.84%) and combined (6.11±7.46%). The between-day analysis in a leave-one-day-out fashion showed that ANN was the optimal classifier (surface (21.88 ± 4.14%) intramuscular (29.33 ± 2.58%) and combined (14.37 ± 3.10%)). Results indicate that that within day performances of classifiers may be similar but over time it may lead to a substantially different outcome. Furthermore, training ANN on multiple days might allow capturing time-dependent variability in the EMG signals and thus minimizing the necessity for daily system recalibration. Index Terms— Electromyography; Pattern recognition; Classification; Myoelectric control; Prostheses; Intramuscular

U2 - 10.1109/JBHI.2018.2864335

DO - 10.1109/JBHI.2018.2864335

M3 - Journal article

VL - 23

SP - 1526

EP - 1534

JO - I E E E Transactions on Information Technology in Biomedicine

JF - I E E E Transactions on Information Technology in Biomedicine

SN - 1089-7771

IS - 4

ER -