Transfer Learning of Spectrogram Image for Automatic Sleep Stage Classification

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

Abstract

Most of the existing methods for automatic sleep stage classification are relying on hand-crafted features. In this paper, the goal is to develop a deep learning-based method that automatically exploits time-frequency spectrum of Electroencephalogram (EEG) signal, removing the need for manual feature extraction. Using Continuous Wavelet Transform (CWT), we extracted the time-frequency spectrogram for EEG signal of 10 healthy subjects and converted to RGB images. The images were classified using transfer learning of a pre-trained Convolutional Neural Network (CNN), AlexNet. The proposed method was evaluated using a publicly available dataset. Evaluation results show that our method can achieve state of the art accuracy, while having higher overall sensitivity.

Original languageEnglish
Title of host publicationImage Analysis and Recognition - 15th International Conference, ICIAR 2018, Proceedings
EditorsB. ter Haar Romeny, F. Karray, A. Campilho
Place of PublicationCham
PublisherSpringer Verlag
Pages522-528
Number of pages7
ISBN (Electronic)978-3-319-93000-8
ISBN (Print)978-3-319-92999-6
DOIs
Publication statusPublished - 1 Jan 2018
Event15th International Conference on Image Analysis and Recognition, ICIAR 2018 - Povoa de Varzim, Portugal
Duration: 27 Jun 201829 Jun 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
PublisherSpringer Verlag
Volume10882 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th International Conference on Image Analysis and Recognition, ICIAR 2018
CountryPortugal
CityPovoa de Varzim
Period27/06/1829/06/18

Fingerprint

Transfer Learning
Spectrogram
Sleep
Electroencephalography
Wavelet transforms
Feature extraction
Neural networks
Continuous Wavelet Transform
Frequency Spectrum
Feature Extraction
Neural Networks
Evaluation
Electroencephalogram
Deep learning

Keywords

  • Convolutional Neural Network
  • Deep learning
  • Discrete Wavelet Transform
  • Sleep stage classification
  • Spectrogram
  • Transfer learning

Cite this

Gharbali, A. A., Najdi, S., & Fonseca, J. M. (2018). Transfer Learning of Spectrogram Image for Automatic Sleep Stage Classification. In B. ter Haar Romeny, F. Karray, & A. Campilho (Eds.), Image Analysis and Recognition - 15th International Conference, ICIAR 2018, Proceedings (pp. 522-528). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10882 LNCS). Cham: Springer Verlag. https://doi.org/10.1007/978-3-319-93000-8_59
Gharbali, Ali Abdollahi ; Najdi, Shirin ; Fonseca, José Manuel. / Transfer Learning of Spectrogram Image for Automatic Sleep Stage Classification. Image Analysis and Recognition - 15th International Conference, ICIAR 2018, Proceedings. editor / B. ter Haar Romeny ; F. Karray ; A. Campilho. Cham : Springer Verlag, 2018. pp. 522-528 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{7df787cbbeaf4fa7b226a311291228cc,
title = "Transfer Learning of Spectrogram Image for Automatic Sleep Stage Classification",
abstract = "Most of the existing methods for automatic sleep stage classification are relying on hand-crafted features. In this paper, the goal is to develop a deep learning-based method that automatically exploits time-frequency spectrum of Electroencephalogram (EEG) signal, removing the need for manual feature extraction. Using Continuous Wavelet Transform (CWT), we extracted the time-frequency spectrogram for EEG signal of 10 healthy subjects and converted to RGB images. The images were classified using transfer learning of a pre-trained Convolutional Neural Network (CNN), AlexNet. The proposed method was evaluated using a publicly available dataset. Evaluation results show that our method can achieve state of the art accuracy, while having higher overall sensitivity.",
keywords = "Convolutional Neural Network, Deep learning, Discrete Wavelet Transform, Sleep stage classification, Spectrogram, Transfer learning",
author = "Gharbali, {Ali Abdollahi} and Shirin Najdi and Fonseca, {Jos{\'e} Manuel}",
note = "UID/EEA/00066/203",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-93000-8_59",
language = "English",
isbn = "978-3-319-92999-6",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "522--528",
editor = "{ter Haar Romeny}, B. and F. Karray and A. Campilho",
booktitle = "Image Analysis and Recognition - 15th International Conference, ICIAR 2018, Proceedings",
address = "Germany",

}

Gharbali, AA, Najdi, S & Fonseca, JM 2018, Transfer Learning of Spectrogram Image for Automatic Sleep Stage Classification. in B ter Haar Romeny, F Karray & A Campilho (eds), Image Analysis and Recognition - 15th International Conference, ICIAR 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10882 LNCS, Springer Verlag, Cham, pp. 522-528, 15th International Conference on Image Analysis and Recognition, ICIAR 2018, Povoa de Varzim, Portugal, 27/06/18. https://doi.org/10.1007/978-3-319-93000-8_59

Transfer Learning of Spectrogram Image for Automatic Sleep Stage Classification. / Gharbali, Ali Abdollahi; Najdi, Shirin; Fonseca, José Manuel.

Image Analysis and Recognition - 15th International Conference, ICIAR 2018, Proceedings. ed. / B. ter Haar Romeny; F. Karray; A. Campilho. Cham : Springer Verlag, 2018. p. 522-528 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10882 LNCS).

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

TY - GEN

T1 - Transfer Learning of Spectrogram Image for Automatic Sleep Stage Classification

AU - Gharbali, Ali Abdollahi

AU - Najdi, Shirin

AU - Fonseca, José Manuel

N1 - UID/EEA/00066/203

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Most of the existing methods for automatic sleep stage classification are relying on hand-crafted features. In this paper, the goal is to develop a deep learning-based method that automatically exploits time-frequency spectrum of Electroencephalogram (EEG) signal, removing the need for manual feature extraction. Using Continuous Wavelet Transform (CWT), we extracted the time-frequency spectrogram for EEG signal of 10 healthy subjects and converted to RGB images. The images were classified using transfer learning of a pre-trained Convolutional Neural Network (CNN), AlexNet. The proposed method was evaluated using a publicly available dataset. Evaluation results show that our method can achieve state of the art accuracy, while having higher overall sensitivity.

AB - Most of the existing methods for automatic sleep stage classification are relying on hand-crafted features. In this paper, the goal is to develop a deep learning-based method that automatically exploits time-frequency spectrum of Electroencephalogram (EEG) signal, removing the need for manual feature extraction. Using Continuous Wavelet Transform (CWT), we extracted the time-frequency spectrogram for EEG signal of 10 healthy subjects and converted to RGB images. The images were classified using transfer learning of a pre-trained Convolutional Neural Network (CNN), AlexNet. The proposed method was evaluated using a publicly available dataset. Evaluation results show that our method can achieve state of the art accuracy, while having higher overall sensitivity.

KW - Convolutional Neural Network

KW - Deep learning

KW - Discrete Wavelet Transform

KW - Sleep stage classification

KW - Spectrogram

KW - Transfer learning

UR - http://www.scopus.com/inward/record.url?scp=85049450698&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-93000-8_59

DO - 10.1007/978-3-319-93000-8_59

M3 - Conference contribution

SN - 978-3-319-92999-6

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

SP - 522

EP - 528

BT - Image Analysis and Recognition - 15th International Conference, ICIAR 2018, Proceedings

A2 - ter Haar Romeny, B.

A2 - Karray, F.

A2 - Campilho, A.

PB - Springer Verlag

CY - Cham

ER -

Gharbali AA, Najdi S, Fonseca JM. Transfer Learning of Spectrogram Image for Automatic Sleep Stage Classification. In ter Haar Romeny B, Karray F, Campilho A, editors, Image Analysis and Recognition - 15th International Conference, ICIAR 2018, Proceedings. Cham: Springer Verlag. 2018. p. 522-528. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-93000-8_59