Skip to content
Back to outputs

Two-dimensional discrete feature based spatial attention CapsNet For sEMG signal recognition

Research output: Contribution to journalArticlepeer-review

Standard

Two-dimensional discrete feature based spatial attention CapsNet For sEMG signal recognition. / Chen, Guoqi; Wang, Wanliang; Wang, Zheng; Liu, Honghai; Zang, Zelin; Li, Weikun.

In: Applied Intelligence, Vol. 50, 01.10.2020, p. 3503-3520.

Research output: Contribution to journalArticlepeer-review

Harvard

Chen, G, Wang, W, Wang, Z, Liu, H, Zang, Z & Li, W 2020, 'Two-dimensional discrete feature based spatial attention CapsNet For sEMG signal recognition', Applied Intelligence, vol. 50, pp. 3503-3520. https://doi.org/10.1007/s10489-020-01725-0

APA

Chen, G., Wang, W., Wang, Z., Liu, H., Zang, Z., & Li, W. (2020). Two-dimensional discrete feature based spatial attention CapsNet For sEMG signal recognition. Applied Intelligence, 50, 3503-3520. https://doi.org/10.1007/s10489-020-01725-0

Vancouver

Author

Chen, Guoqi ; Wang, Wanliang ; Wang, Zheng ; Liu, Honghai ; Zang, Zelin ; Li, Weikun. / Two-dimensional discrete feature based spatial attention CapsNet For sEMG signal recognition. In: Applied Intelligence. 2020 ; Vol. 50. pp. 3503-3520.

Bibtex

@article{1e490c78237d438cb41d861f65963c0a,
title = "Two-dimensional discrete feature based spatial attention CapsNet For sEMG signal recognition",
abstract = "Deep learning frameworks(such as deep convolutional networks) require data to have a regular shape. However, discrete features extracted from heterogeneous data cannot be collected in a regular shape to convolute. In this article, a Two-Dimensional Discrete Feature Based Spatial Attention CapsNet(TDACAPS) is proposed to convert one-dimensional discrete features into two-dimensional structured data through Cartesian Product for surface electromyogram(sEMG) signal recognition. sEMG signal varies from person to person is the main signal source of prosthetic control. Our model transforms multi-angle discrete features into structured data to find the inherent law of sEMG signal. Due to uneven information distribution of structured data, this model combines capsule network with attention mechanism to place emphasis on abundant information regions and reduce ancillary information loss. Extensive experiments show our model yields an improvement for sEMG signal recognition of almost 3% than capsule network and other neural networks under different conditions. Our attention mechanism that employs overlapping pooling to search feature map weight is preferable to the squeeze-and-excitation module, convolutional block attention module and others. Moreover, we validate that our model has great expansibility on Wine Quality Dataset and Breast Cancer Wisconsin.",
author = "Guoqi Chen and Wanliang Wang and Zheng Wang and Honghai Liu and Zelin Zang and Weikun Li",
year = "2020",
month = oct,
day = "1",
doi = "10.1007/s10489-020-01725-0",
language = "English",
volume = "50",
pages = "3503--3520",
journal = "Applied Intelligence",
issn = "0924-669X",
publisher = "Springer",

}

RIS

TY - JOUR

T1 - Two-dimensional discrete feature based spatial attention CapsNet For sEMG signal recognition

AU - Chen, Guoqi

AU - Wang, Wanliang

AU - Wang, Zheng

AU - Liu, Honghai

AU - Zang, Zelin

AU - Li, Weikun

PY - 2020/10/1

Y1 - 2020/10/1

N2 - Deep learning frameworks(such as deep convolutional networks) require data to have a regular shape. However, discrete features extracted from heterogeneous data cannot be collected in a regular shape to convolute. In this article, a Two-Dimensional Discrete Feature Based Spatial Attention CapsNet(TDACAPS) is proposed to convert one-dimensional discrete features into two-dimensional structured data through Cartesian Product for surface electromyogram(sEMG) signal recognition. sEMG signal varies from person to person is the main signal source of prosthetic control. Our model transforms multi-angle discrete features into structured data to find the inherent law of sEMG signal. Due to uneven information distribution of structured data, this model combines capsule network with attention mechanism to place emphasis on abundant information regions and reduce ancillary information loss. Extensive experiments show our model yields an improvement for sEMG signal recognition of almost 3% than capsule network and other neural networks under different conditions. Our attention mechanism that employs overlapping pooling to search feature map weight is preferable to the squeeze-and-excitation module, convolutional block attention module and others. Moreover, we validate that our model has great expansibility on Wine Quality Dataset and Breast Cancer Wisconsin.

AB - Deep learning frameworks(such as deep convolutional networks) require data to have a regular shape. However, discrete features extracted from heterogeneous data cannot be collected in a regular shape to convolute. In this article, a Two-Dimensional Discrete Feature Based Spatial Attention CapsNet(TDACAPS) is proposed to convert one-dimensional discrete features into two-dimensional structured data through Cartesian Product for surface electromyogram(sEMG) signal recognition. sEMG signal varies from person to person is the main signal source of prosthetic control. Our model transforms multi-angle discrete features into structured data to find the inherent law of sEMG signal. Due to uneven information distribution of structured data, this model combines capsule network with attention mechanism to place emphasis on abundant information regions and reduce ancillary information loss. Extensive experiments show our model yields an improvement for sEMG signal recognition of almost 3% than capsule network and other neural networks under different conditions. Our attention mechanism that employs overlapping pooling to search feature map weight is preferable to the squeeze-and-excitation module, convolutional block attention module and others. Moreover, we validate that our model has great expansibility on Wine Quality Dataset and Breast Cancer Wisconsin.

U2 - 10.1007/s10489-020-01725-0

DO - 10.1007/s10489-020-01725-0

M3 - Article

VL - 50

SP - 3503

EP - 3520

JO - Applied Intelligence

JF - Applied Intelligence

SN - 0924-669X

ER -

ID: 21793110