Respiratory sound classification using multi-resolution features

Hyung Jin Seo; Chang Hyeon Jeong; Da Hye Kim; Jong Min Woo; In Chul Yoo; Dongsuk Yook

Respiratory sound classification using multi-resolution features

Hyung Jin Seo, Chang Hyeon Jeong, Da Hye Kim, Jong Min Woo, In Chul Yoo, Dongsuk Yook

Research output: Contribution to journal › Conference article › peer-review

Abstract

With the advance of deep learning technologies, medical image analysis using deep neural networks is achieving very high performance. On the other hand, while sounds contain medical information just as images, medical sound analysis such as automatic identification of adventitious respiratory sound has still been a challenging problem. Most of previous studies for automatic identification of adventitious respiratory sound mainly focused on applying various deep neural networks to the respiratory sound classification using the features developed for automatic speech recognition. In this paper, after the close analysis of adventitious respiratory sounds, we propose to consider a multi-resolution feature to capture the characteristics of widely varying adventitious respiratory sounds. The efficiency of the proposed method is evaluated using the International Conference on Biomedical and Health Informatics (ICBHI) 2017 dataset.

Original language	English
Journal	Proceedings of the International Congress on Acoustics
Publication status	Published - 2022
Event	24th International Congress on Acoustics, ICA 2022 - Gyeongju, Korea, Republic of Duration: 2022 Oct 24 → 2022 Oct 28

Bibliographical note

Publisher Copyright:
© 2022 Proceedings of the International Congress on Acoustics. All rights reserved.

Keywords

Deep neural network
Multi-resolution
Respiratory sound classification

ASJC Scopus subject areas

Mechanical Engineering
Acoustics and Ultrasonics

Cite this

@article{e166398364b04af890b90f793afd9a17,

title = "Respiratory sound classification using multi-resolution features",

abstract = "With the advance of deep learning technologies, medical image analysis using deep neural networks is achieving very high performance. On the other hand, while sounds contain medical information just as images, medical sound analysis such as automatic identification of adventitious respiratory sound has still been a challenging problem. Most of previous studies for automatic identification of adventitious respiratory sound mainly focused on applying various deep neural networks to the respiratory sound classification using the features developed for automatic speech recognition. In this paper, after the close analysis of adventitious respiratory sounds, we propose to consider a multi-resolution feature to capture the characteristics of widely varying adventitious respiratory sounds. The efficiency of the proposed method is evaluated using the International Conference on Biomedical and Health Informatics (ICBHI) 2017 dataset.",

keywords = "Deep neural network, Multi-resolution, Respiratory sound classification",

author = "Seo, {Hyung Jin} and Jeong, {Chang Hyeon} and Kim, {Da Hye} and Woo, {Jong Min} and Yoo, {In Chul} and Dongsuk Yook",

note = "Publisher Copyright: {\textcopyright} 2022 Proceedings of the International Congress on Acoustics. All rights reserved.; 24th International Congress on Acoustics, ICA 2022 ; Conference date: 24-10-2022 Through 28-10-2022",

year = "2022",

language = "English",

journal = "Proceedings of the International Congress on Acoustics",

issn = "2226-7808",

publisher = "International Commission for Acoustics (ICA)",

}

TY - JOUR

T1 - Respiratory sound classification using multi-resolution features

AU - Seo, Hyung Jin

AU - Jeong, Chang Hyeon

AU - Kim, Da Hye

AU - Woo, Jong Min

AU - Yoo, In Chul

AU - Yook, Dongsuk

PY - 2022

Y1 - 2022

N2 - With the advance of deep learning technologies, medical image analysis using deep neural networks is achieving very high performance. On the other hand, while sounds contain medical information just as images, medical sound analysis such as automatic identification of adventitious respiratory sound has still been a challenging problem. Most of previous studies for automatic identification of adventitious respiratory sound mainly focused on applying various deep neural networks to the respiratory sound classification using the features developed for automatic speech recognition. In this paper, after the close analysis of adventitious respiratory sounds, we propose to consider a multi-resolution feature to capture the characteristics of widely varying adventitious respiratory sounds. The efficiency of the proposed method is evaluated using the International Conference on Biomedical and Health Informatics (ICBHI) 2017 dataset.

AB - With the advance of deep learning technologies, medical image analysis using deep neural networks is achieving very high performance. On the other hand, while sounds contain medical information just as images, medical sound analysis such as automatic identification of adventitious respiratory sound has still been a challenging problem. Most of previous studies for automatic identification of adventitious respiratory sound mainly focused on applying various deep neural networks to the respiratory sound classification using the features developed for automatic speech recognition. In this paper, after the close analysis of adventitious respiratory sounds, we propose to consider a multi-resolution feature to capture the characteristics of widely varying adventitious respiratory sounds. The efficiency of the proposed method is evaluated using the International Conference on Biomedical and Health Informatics (ICBHI) 2017 dataset.

KW - Deep neural network

KW - Multi-resolution

KW - Respiratory sound classification

UR - http://www.scopus.com/inward/record.url?scp=85192553507&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85192553507

SN - 2226-7808

JO - Proceedings of the International Congress on Acoustics

JF - Proceedings of the International Congress on Acoustics

T2 - 24th International Congress on Acoustics, ICA 2022

Y2 - 24 October 2022 through 28 October 2022

ER -

Respiratory sound classification using multi-resolution features

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this