Automatic Detection and Classification of Cough Events Based on Deep Learning

Automatic Detection and Classification of Cough Events Based on Deep Learning

In this paper, a deep learning approach for classification of cough sound segments is presented. The architecture of the network is based on a pre-trained network and the spectrogram images of three recording channels have been extracted for the sake of training the network. The classification accur...

Full description

Bibliographic Details
Main Authors: Hossein Tabatabaei Seyed Amir, Augustinov Gabriela, Gross Volker, Sohrabi Keywan, Fischer Patrick, Koehler Ulrich
Format: Article
Language:English
Published: De Gruyter 2020-09-01
Series:Current Directions in Biomedical Engineering
Subjects:
deep learning
convolutional neural networks
respiratory sounds
classification
spectrogram
Online Access:https://doi.org/10.1515/cdbme-2020-3083
id doaj-2be14853ee9d497e84059e0f989aa396
record_format Article
spelling doaj-2be14853ee9d497e84059e0f989aa3962021-09-06T19:19:29ZengDe GruyterCurrent Directions in Biomedical Engineering2364-55042020-09-016332232510.1515/cdbme-2020-3083cdbme-2020-3083Automatic Detection and Classification of Cough Events Based on Deep LearningHossein Tabatabaei Seyed Amir0Augustinov Gabriela1Gross Volker2Sohrabi Keywan3Fischer Patrick4Koehler Ulrich5Institute of Medical Informatics, Justus-Liebig University Giessen,Giessen, GermanyFaculty of Health Sciences, University of Applied Sciences,Giessen, GermanyFaculty of Health Sciences, University of Applied Sciences,Giessen, GermanyFaculty of Health Sciences, University of Applied Sciences,Giessen, GermanyInstitute of Medical Informatics, Justus-Liebig University Giessen,Giessen, GermanyDepartment of Internal Medicine, Pneumology, Intensive Care and Sleep Medicine, University Hospital of Marburg and Giessen,Marburg, GermanyIn this paper, a deep learning approach for classification of cough sound segments is presented. The architecture of the network is based on a pre-trained network and the spectrogram images of three recording channels have been extracted for the sake of training the network. The classification accuracy based on three recording channels is 92% for a binary classification model and the network converges fast. Two classification models based on binary and multi-class problems are proposed. Relevant classification parameters including the Receiver Operating Characteristic (ROC) curve are reported.https://doi.org/10.1515/cdbme-2020-3083deep learningconvolutional neural networksrespiratory soundsclassificationspectrogram
collection DOAJ
language English
format Article
sources DOAJ
author Hossein Tabatabaei Seyed Amir
Augustinov Gabriela
Gross Volker
Sohrabi Keywan
Fischer Patrick
Koehler Ulrich
spellingShingle Hossein Tabatabaei Seyed Amir
Augustinov Gabriela
Gross Volker
Sohrabi Keywan
Fischer Patrick
Koehler Ulrich
Automatic Detection and Classification of Cough Events Based on Deep Learning
Current Directions in Biomedical Engineering
deep learning
convolutional neural networks
respiratory sounds
classification
spectrogram
author_facet Hossein Tabatabaei Seyed Amir
Augustinov Gabriela
Gross Volker
Sohrabi Keywan
Fischer Patrick
Koehler Ulrich
author_sort Hossein Tabatabaei Seyed Amir
title Automatic Detection and Classification of Cough Events Based on Deep Learning
title_short Automatic Detection and Classification of Cough Events Based on Deep Learning
title_full Automatic Detection and Classification of Cough Events Based on Deep Learning
title_fullStr Automatic Detection and Classification of Cough Events Based on Deep Learning
title_full_unstemmed Automatic Detection and Classification of Cough Events Based on Deep Learning
title_sort automatic detection and classification of cough events based on deep learning
publisher De Gruyter
series Current Directions in Biomedical Engineering
issn 2364-5504
publishDate 2020-09-01
description In this paper, a deep learning approach for classification of cough sound segments is presented. The architecture of the network is based on a pre-trained network and the spectrogram images of three recording channels have been extracted for the sake of training the network. The classification accuracy based on three recording channels is 92% for a binary classification model and the network converges fast. Two classification models based on binary and multi-class problems are proposed. Relevant classification parameters including the Receiver Operating Characteristic (ROC) curve are reported.
topic deep learning
convolutional neural networks
respiratory sounds
classification
spectrogram
url https://doi.org/10.1515/cdbme-2020-3083
work_keys_str_mv AT hosseintabatabaeiseyedamir automaticdetectionandclassificationofcougheventsbasedondeeplearning
AT augustinovgabriela automaticdetectionandclassificationofcougheventsbasedondeeplearning
AT grossvolker automaticdetectionandclassificationofcougheventsbasedondeeplearning
AT sohrabikeywan automaticdetectionandclassificationofcougheventsbasedondeeplearning
AT fischerpatrick automaticdetectionandclassificationofcougheventsbasedondeeplearning
AT koehlerulrich automaticdetectionandclassificationofcougheventsbasedondeeplearning
_version_ 1717778475576721408

Similar Items