NeoSSNet: Real-Time Neonatal Chest Sound Separation Using Deep Learning

<italic>Goal:</italic> Auscultation for neonates is a simple and non-invasive method of diagnosing cardiovascular and respiratory disease. However, obtaining high-quality chest sounds containing only heart or lung sounds is non-trivial. Hence, this study introduces a new deep-learning mo...

Full description

Saved in:
Bibliographic Details
Main Authors: Yang Yi Poh, Ethan Grooby, Kenneth Tan, Lindsay Zhou, Arrabella King, Ashwin Ramanathan, Atul Malhotra, Mehrtash Harandi, Faezeh Marzbanrad
Format: Article
Language:English
Published: IEEE 2024-01-01
Series:IEEE Open Journal of Engineering in Medicine and Biology
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10531026/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849422918177521664
author Yang Yi Poh
Ethan Grooby
Kenneth Tan
Lindsay Zhou
Arrabella King
Ashwin Ramanathan
Atul Malhotra
Mehrtash Harandi
Faezeh Marzbanrad
author_facet Yang Yi Poh
Ethan Grooby
Kenneth Tan
Lindsay Zhou
Arrabella King
Ashwin Ramanathan
Atul Malhotra
Mehrtash Harandi
Faezeh Marzbanrad
author_sort Yang Yi Poh
collection DOAJ
description <italic>Goal:</italic> Auscultation for neonates is a simple and non-invasive method of diagnosing cardiovascular and respiratory disease. However, obtaining high-quality chest sounds containing only heart or lung sounds is non-trivial. Hence, this study introduces a new deep-learning model named NeoSSNet and evaluates its performance in neonatal chest sound separation with previous methods. <italic>Methods:</italic> We propose a masked-based architecture similar to Conv-TasNet. The encoder and decoder consist of 1D convolution and 1D transposed convolution, while the mask generator consists of a convolution and transformer architecture. The input chest sounds were first encoded as a sequence of tokens using 1D convolution. The tokens were then passed to the mask generator to generate two masks, one for heart sounds and one for lung sounds. Each mask is then applied to the input token sequence. Lastly, the tokens are converted back to waveforms using 1D transposed convolution. <italic>Results:</italic> Our proposed model showed superior results compared to the previous methods based on objective distortion measures, ranging from a 2.01 dB improvement to a 5.06 dB improvement. The proposed model is also significantly faster than the previous methods, with at least a 17-time improvement. <italic>Conclusions:</italic> The proposed model could be a suitable preprocessing step for any health monitoring system where only the heart sound or lung sound is desired.
format Article
id doaj-art-68eb0d8fbb184e2383757b7c1580dc27
institution Kabale University
issn 2644-1276
language English
publishDate 2024-01-01
publisher IEEE
record_format Article
series IEEE Open Journal of Engineering in Medicine and Biology
spelling doaj-art-68eb0d8fbb184e2383757b7c1580dc272025-08-20T03:30:52ZengIEEEIEEE Open Journal of Engineering in Medicine and Biology2644-12762024-01-01534535210.1109/OJEMB.2024.340157110531026NeoSSNet: Real-Time Neonatal Chest Sound Separation Using Deep LearningYang Yi Poh0https://orcid.org/0009-0001-0869-7047Ethan Grooby1https://orcid.org/0000-0002-1563-4976Kenneth Tan2https://orcid.org/0000-0002-1931-0549Lindsay Zhou3https://orcid.org/0000-0002-3013-5040Arrabella King4Ashwin Ramanathan5https://orcid.org/0000-0002-9298-1234Atul Malhotra6https://orcid.org/0000-0001-9664-4182Mehrtash Harandi7https://orcid.org/0000-0002-6937-6300Faezeh Marzbanrad8https://orcid.org/0000-0003-0551-1611Department of Electrical and Computer Systems Engineering, Monash University, Melbourne, Clayton, VIC, AustraliaDepartment of Electrical and Computer Systems Engineering, Monash University, Melbourne, Clayton, VIC, AustraliaMonash Newborn, Monash Children&#x0027;s Hospital and Department of Paediatrics, Monash University, Melbourne, Clayton, VIC, AustraliaMonash Newborn, Monash Children&#x0027;s Hospital and Department of Paediatrics, Monash University, Melbourne, Clayton, VIC, AustraliaMonash Newborn, Monash Children&#x0027;s Hospital and Department of Paediatrics, Monash University, Melbourne, Clayton, VIC, AustraliaMonash Newborn, Monash Children&#x0027;s Hospital and Department of Paediatrics, Monash University, Melbourne, Clayton, VIC, AustraliaMonash Newborn, Monash Children&#x0027;s Hospital and Department of Paediatrics, Monash University, Melbourne, Clayton, VIC, AustraliaDepartment of Electrical and Computer Systems Engineering, Monash University, Melbourne, Clayton, VIC, AustraliaDepartment of Electrical and Computer Systems Engineering, Monash University, Melbourne, Clayton, VIC, Australia<italic>Goal:</italic> Auscultation for neonates is a simple and non-invasive method of diagnosing cardiovascular and respiratory disease. However, obtaining high-quality chest sounds containing only heart or lung sounds is non-trivial. Hence, this study introduces a new deep-learning model named NeoSSNet and evaluates its performance in neonatal chest sound separation with previous methods. <italic>Methods:</italic> We propose a masked-based architecture similar to Conv-TasNet. The encoder and decoder consist of 1D convolution and 1D transposed convolution, while the mask generator consists of a convolution and transformer architecture. The input chest sounds were first encoded as a sequence of tokens using 1D convolution. The tokens were then passed to the mask generator to generate two masks, one for heart sounds and one for lung sounds. Each mask is then applied to the input token sequence. Lastly, the tokens are converted back to waveforms using 1D transposed convolution. <italic>Results:</italic> Our proposed model showed superior results compared to the previous methods based on objective distortion measures, ranging from a 2.01 dB improvement to a 5.06 dB improvement. The proposed model is also significantly faster than the previous methods, with at least a 17-time improvement. <italic>Conclusions:</italic> The proposed model could be a suitable preprocessing step for any health monitoring system where only the heart sound or lung sound is desired.https://ieeexplore.ieee.org/document/10531026/Deep learningheart soundlung soundphonocardiogram (PCG)single-channel sound separation
spellingShingle Yang Yi Poh
Ethan Grooby
Kenneth Tan
Lindsay Zhou
Arrabella King
Ashwin Ramanathan
Atul Malhotra
Mehrtash Harandi
Faezeh Marzbanrad
NeoSSNet: Real-Time Neonatal Chest Sound Separation Using Deep Learning
IEEE Open Journal of Engineering in Medicine and Biology
Deep learning
heart sound
lung sound
phonocardiogram (PCG)
single-channel sound separation
title NeoSSNet: Real-Time Neonatal Chest Sound Separation Using Deep Learning
title_full NeoSSNet: Real-Time Neonatal Chest Sound Separation Using Deep Learning
title_fullStr NeoSSNet: Real-Time Neonatal Chest Sound Separation Using Deep Learning
title_full_unstemmed NeoSSNet: Real-Time Neonatal Chest Sound Separation Using Deep Learning
title_short NeoSSNet: Real-Time Neonatal Chest Sound Separation Using Deep Learning
title_sort neossnet real time neonatal chest sound separation using deep learning
topic Deep learning
heart sound
lung sound
phonocardiogram (PCG)
single-channel sound separation
url https://ieeexplore.ieee.org/document/10531026/
work_keys_str_mv AT yangyipoh neossnetrealtimeneonatalchestsoundseparationusingdeeplearning
AT ethangrooby neossnetrealtimeneonatalchestsoundseparationusingdeeplearning
AT kennethtan neossnetrealtimeneonatalchestsoundseparationusingdeeplearning
AT lindsayzhou neossnetrealtimeneonatalchestsoundseparationusingdeeplearning
AT arrabellaking neossnetrealtimeneonatalchestsoundseparationusingdeeplearning
AT ashwinramanathan neossnetrealtimeneonatalchestsoundseparationusingdeeplearning
AT atulmalhotra neossnetrealtimeneonatalchestsoundseparationusingdeeplearning
AT mehrtashharandi neossnetrealtimeneonatalchestsoundseparationusingdeeplearning
AT faezehmarzbanrad neossnetrealtimeneonatalchestsoundseparationusingdeeplearning