A machine hearing system for robust cough detection based on a high-level representation of band-specific audio features

Research output: Contribution to journalArticle

21 Downloads (Pure)

Abstract

Cough is a protective reflex conveying information on the state of the respiratory system. Cough assessment has been limited so far to subjective measurement tools or uncomfortable (i.e., non-wearable) cough monitors. This limits the
potential of real-time cough monitoring to improve respiratory care.

Objective: This paper presents a machine hearing system for audio-based robust cough segmentation that can be easily deployed in mobile scenarios.

Methods: Cough detection is performed in two steps. First, a short-term spectral feature set is separately computed in five pre-defined frequency bands: [0,0.5), [0.5,1), [1,1.5), [1.5,2), [2,5.5125] kHz. Feature selection and combination are then applied to make the short-term feature set robust enough in different noisy scenarios. Secondly, high level data representation is achieved by computing the mean and standard deviation of short-term descriptors in 300 ms
long-term frames. Finally, cough detection is carried out using a support vector machine trained with data from different noisy scenarios. The system is evaluated using a patient signal database which emulates three real-life scenarios in terms of noise content.

Results: the system achieves 92.71% sensitivity, 88.58% specificity, and 90.69% Area Under Receiver Operating Characteristic (ROC) curve (AUC), outperforming state-of-the art methods, outperforming state-of-the-art methods.

Conclusion: our research outcome paves the way to create a device for cough monitoring in real-life situations. Significance: our proposal is aligned with a more comfortable and less disruptive patient monitoring, with benefits for patients (allows self-monitoring of cough symptoms), practitioners (e.g., assessment of treatments or better clinical understanding of cough patterns) and national health systems (by reducing hospitalisations).
Original languageEnglish
Pages (from-to)2319-2330
Number of pages12
JournalIEEE Transactions on Biomedical Engineering
Volume66
Issue number8
Early online date20 Dec 2018
DOIs
Publication statusPublished - 31 Aug 2019

Fingerprint

Audition
Monitoring
Respiratory system
Patient monitoring
Conveying
Frequency bands
Support vector machines
Feature extraction
Health

Keywords

  • Cough detection
  • machine hearing
  • respiratory care
  • patient monitoring
  • spectral features

Cite this

@article{2196248b05d146ed895c773c08667475,
title = "A machine hearing system for robust cough detection based on a high-level representation of band-specific audio features",
abstract = "Cough is a protective reflex conveying information on the state of the respiratory system. Cough assessment has been limited so far to subjective measurement tools or uncomfortable (i.e., non-wearable) cough monitors. This limits thepotential of real-time cough monitoring to improve respiratory care. Objective: This paper presents a machine hearing system for audio-based robust cough segmentation that can be easily deployed in mobile scenarios. Methods: Cough detection is performed in two steps. First, a short-term spectral feature set is separately computed in five pre-defined frequency bands: [0,0.5), [0.5,1), [1,1.5), [1.5,2), [2,5.5125] kHz. Feature selection and combination are then applied to make the short-term feature set robust enough in different noisy scenarios. Secondly, high level data representation is achieved by computing the mean and standard deviation of short-term descriptors in 300 mslong-term frames. Finally, cough detection is carried out using a support vector machine trained with data from different noisy scenarios. The system is evaluated using a patient signal database which emulates three real-life scenarios in terms of noise content. Results: the system achieves 92.71{\%} sensitivity, 88.58{\%} specificity, and 90.69{\%} Area Under Receiver Operating Characteristic (ROC) curve (AUC), outperforming state-of-the art methods, outperforming state-of-the-art methods. Conclusion: our research outcome paves the way to create a device for cough monitoring in real-life situations. Significance: our proposal is aligned with a more comfortable and less disruptive patient monitoring, with benefits for patients (allows self-monitoring of cough symptoms), practitioners (e.g., assessment of treatments or better clinical understanding of cough patterns) and national health systems (by reducing hospitalisations).",
keywords = "Cough detection, machine hearing, respiratory care, patient monitoring, spectral features",
author = "Jes{\'u}s Monge-{\'A}lvarez and {Hoyos Barcel{\'o}}, Carlos and San-Jos{\'e}-Revuelta, {Luis M.} and Pablo Casaseca-de-la-Higuera",
year = "2019",
month = "8",
day = "31",
doi = "10.1109/TBME.2018.2888998",
language = "English",
volume = "66",
pages = "2319--2330",
journal = "IEEE Transactions on Biomedical Engineering",
issn = "0018-9294",
publisher = "IEEE",
number = "8",

}

TY - JOUR

T1 - A machine hearing system for robust cough detection based on a high-level representation of band-specific audio features

AU - Monge-Álvarez, Jesús

AU - Hoyos Barceló, Carlos

AU - San-José-Revuelta, Luis M.

AU - Casaseca-de-la-Higuera, Pablo

PY - 2019/8/31

Y1 - 2019/8/31

N2 - Cough is a protective reflex conveying information on the state of the respiratory system. Cough assessment has been limited so far to subjective measurement tools or uncomfortable (i.e., non-wearable) cough monitors. This limits thepotential of real-time cough monitoring to improve respiratory care. Objective: This paper presents a machine hearing system for audio-based robust cough segmentation that can be easily deployed in mobile scenarios. Methods: Cough detection is performed in two steps. First, a short-term spectral feature set is separately computed in five pre-defined frequency bands: [0,0.5), [0.5,1), [1,1.5), [1.5,2), [2,5.5125] kHz. Feature selection and combination are then applied to make the short-term feature set robust enough in different noisy scenarios. Secondly, high level data representation is achieved by computing the mean and standard deviation of short-term descriptors in 300 mslong-term frames. Finally, cough detection is carried out using a support vector machine trained with data from different noisy scenarios. The system is evaluated using a patient signal database which emulates three real-life scenarios in terms of noise content. Results: the system achieves 92.71% sensitivity, 88.58% specificity, and 90.69% Area Under Receiver Operating Characteristic (ROC) curve (AUC), outperforming state-of-the art methods, outperforming state-of-the-art methods. Conclusion: our research outcome paves the way to create a device for cough monitoring in real-life situations. Significance: our proposal is aligned with a more comfortable and less disruptive patient monitoring, with benefits for patients (allows self-monitoring of cough symptoms), practitioners (e.g., assessment of treatments or better clinical understanding of cough patterns) and national health systems (by reducing hospitalisations).

AB - Cough is a protective reflex conveying information on the state of the respiratory system. Cough assessment has been limited so far to subjective measurement tools or uncomfortable (i.e., non-wearable) cough monitors. This limits thepotential of real-time cough monitoring to improve respiratory care. Objective: This paper presents a machine hearing system for audio-based robust cough segmentation that can be easily deployed in mobile scenarios. Methods: Cough detection is performed in two steps. First, a short-term spectral feature set is separately computed in five pre-defined frequency bands: [0,0.5), [0.5,1), [1,1.5), [1.5,2), [2,5.5125] kHz. Feature selection and combination are then applied to make the short-term feature set robust enough in different noisy scenarios. Secondly, high level data representation is achieved by computing the mean and standard deviation of short-term descriptors in 300 mslong-term frames. Finally, cough detection is carried out using a support vector machine trained with data from different noisy scenarios. The system is evaluated using a patient signal database which emulates three real-life scenarios in terms of noise content. Results: the system achieves 92.71% sensitivity, 88.58% specificity, and 90.69% Area Under Receiver Operating Characteristic (ROC) curve (AUC), outperforming state-of-the art methods, outperforming state-of-the-art methods. Conclusion: our research outcome paves the way to create a device for cough monitoring in real-life situations. Significance: our proposal is aligned with a more comfortable and less disruptive patient monitoring, with benefits for patients (allows self-monitoring of cough symptoms), practitioners (e.g., assessment of treatments or better clinical understanding of cough patterns) and national health systems (by reducing hospitalisations).

KW - Cough detection

KW - machine hearing

KW - respiratory care

KW - patient monitoring

KW - spectral features

U2 - 10.1109/TBME.2018.2888998

DO - 10.1109/TBME.2018.2888998

M3 - Article

VL - 66

SP - 2319

EP - 2330

JO - IEEE Transactions on Biomedical Engineering

JF - IEEE Transactions on Biomedical Engineering

SN - 0018-9294

IS - 8

ER -