Subject-independent hand gesture recognition using normalization and machine learning algorithms

Md Ferdous Wahid, Reza Tafreshi, Mubarak Al-Sowaidi, Reza Langari

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Hand gestures can be recognized using the upper limb's electromyography (EMG) that measures the electrical activity of the skeletal muscles. However, generalization of muscle activities for a particular hand gesture is challenging due to between-subject variations in EMG signals. To improve the gesture recognition accuracy without training the machine learning algorithm subject specifically, the time-domain EMG features are normalized to the area under the averaged root mean square curve (AUC-RMS). Results are compared with both original EMG features and EMG features extracted from the signals that are normalized to the maximum peak value. Ten male adult subjects age ranging 20–37 years performed three hand gestures including fist, wave in, and wave out for ten to twelve times. The four basic time domain features including mean absolute value, zero crossing, waveform length, and slope sign change were extracted from the active EMG signals of each channel. Five machine learning algorithms, namely, k-Nearest Neighbor (kNN), Discriminant Analysis (DA), Naïve Bayes (NB), Random Forest (RF), and Support Vector Machine (SVM) were used to classify the three different hand gestures. The results showed that the performance metrics such as accuracy, F1-score, Matthew correlation coefficient, and Kappa score were improved when using the both normalization methods compared to the original EMG features. However, normalization to the AUC-RMS value resulted in substantially more accurate gesture recognition compared to features extracted from signal normalized to maximum peak value using kNN, NB, and RF (p < 0.05). The developed approach of classifying different hand gestures will be useful in human-computer interaction as well as in controlling devices including prosthesis, virtual objects, and wheelchair.

Original languageEnglish
Pages (from-to)69-76
Number of pages8
JournalJournal of Computational Science
Volume27
DOIs
Publication statusPublished - 1 Jul 2018

Fingerprint

Hand Gesture Recognition
Electromyography
Gesture recognition
Gesture
Learning algorithms
Normalization
Learning systems
Learning Algorithm
Machine Learning
Gesture Recognition
Random Forest
Bayes
Time Domain
Nearest Neighbor
Zero-crossing
Muscle
Skeletal muscle
Sign Change
Performance Metrics
Discriminant Analysis

Keywords

  • Electromyography
  • Hand gesture
  • Machine learning
  • MYO armband
  • Normalization
  • Pattern recognition
  • Time-domain

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)
  • Modelling and Simulation

Cite this

Subject-independent hand gesture recognition using normalization and machine learning algorithms. / Wahid, Md Ferdous; Tafreshi, Reza; Al-Sowaidi, Mubarak; Langari, Reza.

In: Journal of Computational Science, Vol. 27, 01.07.2018, p. 69-76.

Research output: Contribution to journalArticle

@article{e5236c55f70e40b4a3f643a35c7569c5,
title = "Subject-independent hand gesture recognition using normalization and machine learning algorithms",
abstract = "Hand gestures can be recognized using the upper limb's electromyography (EMG) that measures the electrical activity of the skeletal muscles. However, generalization of muscle activities for a particular hand gesture is challenging due to between-subject variations in EMG signals. To improve the gesture recognition accuracy without training the machine learning algorithm subject specifically, the time-domain EMG features are normalized to the area under the averaged root mean square curve (AUC-RMS). Results are compared with both original EMG features and EMG features extracted from the signals that are normalized to the maximum peak value. Ten male adult subjects age ranging 20–37 years performed three hand gestures including fist, wave in, and wave out for ten to twelve times. The four basic time domain features including mean absolute value, zero crossing, waveform length, and slope sign change were extracted from the active EMG signals of each channel. Five machine learning algorithms, namely, k-Nearest Neighbor (kNN), Discriminant Analysis (DA), Na{\"i}ve Bayes (NB), Random Forest (RF), and Support Vector Machine (SVM) were used to classify the three different hand gestures. The results showed that the performance metrics such as accuracy, F1-score, Matthew correlation coefficient, and Kappa score were improved when using the both normalization methods compared to the original EMG features. However, normalization to the AUC-RMS value resulted in substantially more accurate gesture recognition compared to features extracted from signal normalized to maximum peak value using kNN, NB, and RF (p < 0.05). The developed approach of classifying different hand gestures will be useful in human-computer interaction as well as in controlling devices including prosthesis, virtual objects, and wheelchair.",
keywords = "Electromyography, Hand gesture, Machine learning, MYO armband, Normalization, Pattern recognition, Time-domain",
author = "Wahid, {Md Ferdous} and Reza Tafreshi and Mubarak Al-Sowaidi and Reza Langari",
year = "2018",
month = "7",
day = "1",
doi = "10.1016/j.jocs.2018.04.019",
language = "English",
volume = "27",
pages = "69--76",
journal = "Journal of Computational Science",
issn = "1877-7503",
publisher = "Elsevier",

}

TY - JOUR

T1 - Subject-independent hand gesture recognition using normalization and machine learning algorithms

AU - Wahid, Md Ferdous

AU - Tafreshi, Reza

AU - Al-Sowaidi, Mubarak

AU - Langari, Reza

PY - 2018/7/1

Y1 - 2018/7/1

N2 - Hand gestures can be recognized using the upper limb's electromyography (EMG) that measures the electrical activity of the skeletal muscles. However, generalization of muscle activities for a particular hand gesture is challenging due to between-subject variations in EMG signals. To improve the gesture recognition accuracy without training the machine learning algorithm subject specifically, the time-domain EMG features are normalized to the area under the averaged root mean square curve (AUC-RMS). Results are compared with both original EMG features and EMG features extracted from the signals that are normalized to the maximum peak value. Ten male adult subjects age ranging 20–37 years performed three hand gestures including fist, wave in, and wave out for ten to twelve times. The four basic time domain features including mean absolute value, zero crossing, waveform length, and slope sign change were extracted from the active EMG signals of each channel. Five machine learning algorithms, namely, k-Nearest Neighbor (kNN), Discriminant Analysis (DA), Naïve Bayes (NB), Random Forest (RF), and Support Vector Machine (SVM) were used to classify the three different hand gestures. The results showed that the performance metrics such as accuracy, F1-score, Matthew correlation coefficient, and Kappa score were improved when using the both normalization methods compared to the original EMG features. However, normalization to the AUC-RMS value resulted in substantially more accurate gesture recognition compared to features extracted from signal normalized to maximum peak value using kNN, NB, and RF (p < 0.05). The developed approach of classifying different hand gestures will be useful in human-computer interaction as well as in controlling devices including prosthesis, virtual objects, and wheelchair.

AB - Hand gestures can be recognized using the upper limb's electromyography (EMG) that measures the electrical activity of the skeletal muscles. However, generalization of muscle activities for a particular hand gesture is challenging due to between-subject variations in EMG signals. To improve the gesture recognition accuracy without training the machine learning algorithm subject specifically, the time-domain EMG features are normalized to the area under the averaged root mean square curve (AUC-RMS). Results are compared with both original EMG features and EMG features extracted from the signals that are normalized to the maximum peak value. Ten male adult subjects age ranging 20–37 years performed three hand gestures including fist, wave in, and wave out for ten to twelve times. The four basic time domain features including mean absolute value, zero crossing, waveform length, and slope sign change were extracted from the active EMG signals of each channel. Five machine learning algorithms, namely, k-Nearest Neighbor (kNN), Discriminant Analysis (DA), Naïve Bayes (NB), Random Forest (RF), and Support Vector Machine (SVM) were used to classify the three different hand gestures. The results showed that the performance metrics such as accuracy, F1-score, Matthew correlation coefficient, and Kappa score were improved when using the both normalization methods compared to the original EMG features. However, normalization to the AUC-RMS value resulted in substantially more accurate gesture recognition compared to features extracted from signal normalized to maximum peak value using kNN, NB, and RF (p < 0.05). The developed approach of classifying different hand gestures will be useful in human-computer interaction as well as in controlling devices including prosthesis, virtual objects, and wheelchair.

KW - Electromyography

KW - Hand gesture

KW - Machine learning

KW - MYO armband

KW - Normalization

KW - Pattern recognition

KW - Time-domain

UR - http://www.scopus.com/inward/record.url?scp=85047092579&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047092579&partnerID=8YFLogxK

U2 - 10.1016/j.jocs.2018.04.019

DO - 10.1016/j.jocs.2018.04.019

M3 - Article

AN - SCOPUS:85047092579

VL - 27

SP - 69

EP - 76

JO - Journal of Computational Science

JF - Journal of Computational Science

SN - 1877-7503

ER -