Motor imagery task classification using a signal-dependent orthogonal transform based feature extraction

Mostefa Mesbah, Aida Khorshidtalab, Hamza Baali, Ahmed Al-Ani

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

In this paper, we present the results of classifying electroencephalographic (EEG) signals into four motor imagery tasks using a new method for feature extraction. This method is based on a signal-dependent orthogonal transform, referred to as LP-SVD, defined as the left singular vectors of the LPC filter impulse response matrix. Using a logistic tree based model classifier, the extracted features are mapped into one of four motor imagery movements, namely left hand, right hand, foot, and tongue. The proposed technique-based classification performance was benchmarked against those based on two widely used linear transform for feature extraction methods, namely discrete cosine transform (DCT) and adaptive autoregressive (AAR). By achieving an accuracy of 67.35 %, the LP-SVD based method outperformed the other two by large margins (+25 % compared to DCT and +6 % compared to AAR-based methods).

Original languageEnglish
Title of host publicationNeural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings
PublisherSpringer Verlag
Pages1-9
Number of pages9
Volume9490
ISBN (Print)9783319265346
DOIs
Publication statusPublished - 2015
Event22nd International Conference on Neural Information Processing, ICONIP 2015 - Istanbul, Turkey
Duration: Nov 9 2015Nov 12 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9490
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other22nd International Conference on Neural Information Processing, ICONIP 2015
CountryTurkey
CityIstanbul
Period11/9/1511/12/15

Fingerprint

Discrete cosine transforms
Singular value decomposition
Feature Extraction
Feature extraction
Transform
Dependent
Impulse response
Discrete Cosine Transform
Logistics
Classifiers
Singular Vectors
Impulse Response
Margin
Classifier
Imagery
Filter

Keywords

  • Brain-computer interface
  • Feature extraction
  • Linear prediction
  • Orthogonal transform
  • SVD

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Mesbah, M., Khorshidtalab, A., Baali, H., & Al-Ani, A. (2015). Motor imagery task classification using a signal-dependent orthogonal transform based feature extraction. In Neural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings (Vol. 9490, pp. 1-9). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9490). Springer Verlag. https://doi.org/10.1007/978-3-319-26535-3_1

Motor imagery task classification using a signal-dependent orthogonal transform based feature extraction. / Mesbah, Mostefa; Khorshidtalab, Aida; Baali, Hamza; Al-Ani, Ahmed.

Neural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings. Vol. 9490 Springer Verlag, 2015. p. 1-9 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9490).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Mesbah, M, Khorshidtalab, A, Baali, H & Al-Ani, A 2015, Motor imagery task classification using a signal-dependent orthogonal transform based feature extraction. in Neural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings. vol. 9490, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9490, Springer Verlag, pp. 1-9, 22nd International Conference on Neural Information Processing, ICONIP 2015, Istanbul, Turkey, 11/9/15. https://doi.org/10.1007/978-3-319-26535-3_1
Mesbah M, Khorshidtalab A, Baali H, Al-Ani A. Motor imagery task classification using a signal-dependent orthogonal transform based feature extraction. In Neural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings. Vol. 9490. Springer Verlag. 2015. p. 1-9. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-26535-3_1
Mesbah, Mostefa ; Khorshidtalab, Aida ; Baali, Hamza ; Al-Ani, Ahmed. / Motor imagery task classification using a signal-dependent orthogonal transform based feature extraction. Neural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings. Vol. 9490 Springer Verlag, 2015. pp. 1-9 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{7b75829efd894100a9bcfdd9820832db,
title = "Motor imagery task classification using a signal-dependent orthogonal transform based feature extraction",
abstract = "In this paper, we present the results of classifying electroencephalographic (EEG) signals into four motor imagery tasks using a new method for feature extraction. This method is based on a signal-dependent orthogonal transform, referred to as LP-SVD, defined as the left singular vectors of the LPC filter impulse response matrix. Using a logistic tree based model classifier, the extracted features are mapped into one of four motor imagery movements, namely left hand, right hand, foot, and tongue. The proposed technique-based classification performance was benchmarked against those based on two widely used linear transform for feature extraction methods, namely discrete cosine transform (DCT) and adaptive autoregressive (AAR). By achieving an accuracy of 67.35 {\%}, the LP-SVD based method outperformed the other two by large margins (+25 {\%} compared to DCT and +6 {\%} compared to AAR-based methods).",
keywords = "Brain-computer interface, Feature extraction, Linear prediction, Orthogonal transform, SVD",
author = "Mostefa Mesbah and Aida Khorshidtalab and Hamza Baali and Ahmed Al-Ani",
year = "2015",
doi = "10.1007/978-3-319-26535-3_1",
language = "English",
isbn = "9783319265346",
volume = "9490",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "1--9",
booktitle = "Neural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings",

}

TY - GEN

T1 - Motor imagery task classification using a signal-dependent orthogonal transform based feature extraction

AU - Mesbah, Mostefa

AU - Khorshidtalab, Aida

AU - Baali, Hamza

AU - Al-Ani, Ahmed

PY - 2015

Y1 - 2015

N2 - In this paper, we present the results of classifying electroencephalographic (EEG) signals into four motor imagery tasks using a new method for feature extraction. This method is based on a signal-dependent orthogonal transform, referred to as LP-SVD, defined as the left singular vectors of the LPC filter impulse response matrix. Using a logistic tree based model classifier, the extracted features are mapped into one of four motor imagery movements, namely left hand, right hand, foot, and tongue. The proposed technique-based classification performance was benchmarked against those based on two widely used linear transform for feature extraction methods, namely discrete cosine transform (DCT) and adaptive autoregressive (AAR). By achieving an accuracy of 67.35 %, the LP-SVD based method outperformed the other two by large margins (+25 % compared to DCT and +6 % compared to AAR-based methods).

AB - In this paper, we present the results of classifying electroencephalographic (EEG) signals into four motor imagery tasks using a new method for feature extraction. This method is based on a signal-dependent orthogonal transform, referred to as LP-SVD, defined as the left singular vectors of the LPC filter impulse response matrix. Using a logistic tree based model classifier, the extracted features are mapped into one of four motor imagery movements, namely left hand, right hand, foot, and tongue. The proposed technique-based classification performance was benchmarked against those based on two widely used linear transform for feature extraction methods, namely discrete cosine transform (DCT) and adaptive autoregressive (AAR). By achieving an accuracy of 67.35 %, the LP-SVD based method outperformed the other two by large margins (+25 % compared to DCT and +6 % compared to AAR-based methods).

KW - Brain-computer interface

KW - Feature extraction

KW - Linear prediction

KW - Orthogonal transform

KW - SVD

UR - http://www.scopus.com/inward/record.url?scp=84951750598&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84951750598&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-26535-3_1

DO - 10.1007/978-3-319-26535-3_1

M3 - Conference contribution

SN - 9783319265346

VL - 9490

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 1

EP - 9

BT - Neural Information Processing - 22nd International Conference, ICONIP 2015, Proceedings

PB - Springer Verlag

ER -