Analysis and optimization of novel post-processing method for myoelectric pattern recognition

Masahiro Kasuya, Hiroshi Yokoi, Ryu Kato

    Abstract

    This paper describes a novel post-processing step for an electromyogram (EMG) pattern classification algorithm that is used in the control of myoelectric prosthetic hands. Amputees often find it difficult to control multiple degrees of freedom, but increasing numbers of prosthetic hands have multiple degrees of freedom. In general, larger numbers of classes tend to reduce the classification accuracy, and artificial neural networks have been used in previous studies for EMG pattern classification. The proposed post-processing algorithm stores the temporal sequence of classifications from the EMG pattern classification algorithm, and then runs a second classification based on the sequential patterns. We compared the output accuracy before and after the post-processing step. In our experiment, we set the training time for the EMG pattern classification algorithm to 1 s for each class, and used three channels of surface EMG signals. We considered nine EMG pattern classes, and recorded the output every 10-20 ms. We then analyzed the proposed algorithm and found seven classes to be the number required for optimal performance. The overall accuracy of the proposed system was 90.0% for seven classes and 89.8% for nine classes. The classification accuracy improved by 12.5% when using seven classes, and by 21.2% when using nine classes. We believe that this level of classification accuracy and other elements (the number of EMG channels and the training time) are sufficient for practical use with prosthetic hands.

    Original languageEnglish
    Title of host publicationIEEE International Conference on Rehabilitation Robotics
    PublisherIEEE Computer Society
    Pages985-990
    Number of pages6
    Volume2015-September
    ISBN (Print)9781479918072
    DOIs
    StatePublished - 2015 Sep 28
    Event14th IEEE/RAS-EMBS International Conference on Rehabilitation Robotics, ICORR 2015 - Singapore, Singapore

    Other

    Other14th IEEE/RAS-EMBS International Conference on Rehabilitation Robotics, ICORR 2015
    CountrySingapore
    CitySingapore
    Period15/8/1115/8/14

    Fingerprint

    Pattern recognition
    Prosthetics
    Neural networks
    Experiments

    Keywords

    • human-machine interaction
    • prosthetic devices
    • rehabilitation and assistive robotics

    ASJC Scopus subject areas

    • Control and Systems Engineering
    • Electrical and Electronic Engineering
    • Rehabilitation

    Cite this

    Kasuya, M., Yokoi, H., & Kato, R. (2015). Analysis and optimization of novel post-processing method for myoelectric pattern recognition. In IEEE International Conference on Rehabilitation Robotics. (Vol. 2015-September, pp. 985-990). [7281332] IEEE Computer Society. DOI: 10.1109/ICORR.2015.7281332

    Analysis and optimization of novel post-processing method for myoelectric pattern recognition. / Kasuya, Masahiro; Yokoi, Hiroshi; Kato, Ryu.

    IEEE International Conference on Rehabilitation Robotics. Vol. 2015-September IEEE Computer Society, 2015. p. 985-990 7281332.

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

    Kasuya, M, Yokoi, H & Kato, R 2015, Analysis and optimization of novel post-processing method for myoelectric pattern recognition. in IEEE International Conference on Rehabilitation Robotics. vol. 2015-September, 7281332, IEEE Computer Society, pp. 985-990, 14th IEEE/RAS-EMBS International Conference on Rehabilitation Robotics, ICORR 2015, Singapore, Singapore, 11-14 August. DOI: 10.1109/ICORR.2015.7281332
    Kasuya M, Yokoi H, Kato R. Analysis and optimization of novel post-processing method for myoelectric pattern recognition. In IEEE International Conference on Rehabilitation Robotics. Vol. 2015-September. IEEE Computer Society. 2015. p. 985-990. 7281332. Available from, DOI: 10.1109/ICORR.2015.7281332

    Kasuya, Masahiro; Yokoi, Hiroshi; Kato, Ryu / Analysis and optimization of novel post-processing method for myoelectric pattern recognition.

    IEEE International Conference on Rehabilitation Robotics. Vol. 2015-September IEEE Computer Society, 2015. p. 985-990 7281332.

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

    @inbook{ec27fdbbe792435aa512ca51af44e9e0,
    title = "Analysis and optimization of novel post-processing method for myoelectric pattern recognition",
    keywords = "human-machine interaction, prosthetic devices, rehabilitation and assistive robotics",
    author = "Masahiro Kasuya and Hiroshi Yokoi and Ryu Kato",
    year = "2015",
    month = "9",
    doi = "10.1109/ICORR.2015.7281332",
    isbn = "9781479918072",
    volume = "2015-September",
    pages = "985--990",
    booktitle = "IEEE International Conference on Rehabilitation Robotics",
    publisher = "IEEE Computer Society",

    }

    TY - CHAP

    T1 - Analysis and optimization of novel post-processing method for myoelectric pattern recognition

    AU - Kasuya,Masahiro

    AU - Yokoi,Hiroshi

    AU - Kato,Ryu

    PY - 2015/9/28

    Y1 - 2015/9/28

    N2 - This paper describes a novel post-processing step for an electromyogram (EMG) pattern classification algorithm that is used in the control of myoelectric prosthetic hands. Amputees often find it difficult to control multiple degrees of freedom, but increasing numbers of prosthetic hands have multiple degrees of freedom. In general, larger numbers of classes tend to reduce the classification accuracy, and artificial neural networks have been used in previous studies for EMG pattern classification. The proposed post-processing algorithm stores the temporal sequence of classifications from the EMG pattern classification algorithm, and then runs a second classification based on the sequential patterns. We compared the output accuracy before and after the post-processing step. In our experiment, we set the training time for the EMG pattern classification algorithm to 1 s for each class, and used three channels of surface EMG signals. We considered nine EMG pattern classes, and recorded the output every 10-20 ms. We then analyzed the proposed algorithm and found seven classes to be the number required for optimal performance. The overall accuracy of the proposed system was 90.0% for seven classes and 89.8% for nine classes. The classification accuracy improved by 12.5% when using seven classes, and by 21.2% when using nine classes. We believe that this level of classification accuracy and other elements (the number of EMG channels and the training time) are sufficient for practical use with prosthetic hands.

    AB - This paper describes a novel post-processing step for an electromyogram (EMG) pattern classification algorithm that is used in the control of myoelectric prosthetic hands. Amputees often find it difficult to control multiple degrees of freedom, but increasing numbers of prosthetic hands have multiple degrees of freedom. In general, larger numbers of classes tend to reduce the classification accuracy, and artificial neural networks have been used in previous studies for EMG pattern classification. The proposed post-processing algorithm stores the temporal sequence of classifications from the EMG pattern classification algorithm, and then runs a second classification based on the sequential patterns. We compared the output accuracy before and after the post-processing step. In our experiment, we set the training time for the EMG pattern classification algorithm to 1 s for each class, and used three channels of surface EMG signals. We considered nine EMG pattern classes, and recorded the output every 10-20 ms. We then analyzed the proposed algorithm and found seven classes to be the number required for optimal performance. The overall accuracy of the proposed system was 90.0% for seven classes and 89.8% for nine classes. The classification accuracy improved by 12.5% when using seven classes, and by 21.2% when using nine classes. We believe that this level of classification accuracy and other elements (the number of EMG channels and the training time) are sufficient for practical use with prosthetic hands.

    KW - human-machine interaction

    KW - prosthetic devices

    KW - rehabilitation and assistive robotics

    UR - http://www.scopus.com/inward/record.url?scp=84946064417&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84946064417&partnerID=8YFLogxK

    U2 - 10.1109/ICORR.2015.7281332

    DO - 10.1109/ICORR.2015.7281332

    M3 - Conference contribution

    SN - 9781479918072

    VL - 2015-September

    SP - 985

    EP - 990

    BT - IEEE International Conference on Rehabilitation Robotics

    PB - IEEE Computer Society

    ER -