Efficient ECG-based Atrial Fibrillation Detection via Parameterised Hypercomplex Neural Networks

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Autoren

Organisationseinheiten

Forschungs-netzwerk anzeigen

Details

OriginalspracheEnglisch
Titel des Sammelwerks31st European Signal Processing Conference
UntertitelEUSIPCO 2023
Seiten1375-1379
Seitenumfang5
ISBN (elektronisch)9789464593600
PublikationsstatusVeröffentlicht - 2023
Veranstaltung31st European Signal Processing Conference, EUSIPCO 2023 - Helsinki, Finnland
Dauer: 4 Sept. 20238 Sept. 2023

Publikationsreihe

NameEuropean Signal Processing Conference
ISSN (Print)2219-5491

Abstract

Atrial fibrillation (AF) is the most common cardiac arrhythmia and associated with a high risk for serious conditions like stroke. The use of wearable devices embedded with automatic and timely AF assessment from electrocardiograms (ECGs) has shown to be promising in preventing life-threatening situations. Although deep neural networks have demonstrated superiority in model performance, their use on wearable devices is limited by the trade-off between model performance and complexity. In this work, we propose to use lightweight convolutional neural networks (CNNs) with parameterised hypercomplex (PH) layers for AF detection based on ECGs. The proposed approach trains small-scale CNNs, thus overcoming the limited computing resources on wearable devices. We show comparable performance to corresponding real-valued CNNs on two publicly available ECG datasets using significantly fewer model parameters. PH models are more flexible than other hypercomplex neural networks and can operate on any number of input ECG leads.

ASJC Scopus Sachgebiete

Zitieren

Efficient ECG-based Atrial Fibrillation Detection via Parameterised Hypercomplex Neural Networks. / Basso, Leonie; Ren, Zhao; Nejdl, Wolfgang.
31st European Signal Processing Conference: EUSIPCO 2023. 2023. S. 1375-1379 (European Signal Processing Conference).

Publikation: Beitrag in Buch/Bericht/Sammelwerk/KonferenzbandAufsatz in KonferenzbandForschungPeer-Review

Basso, L, Ren, Z & Nejdl, W 2023, Efficient ECG-based Atrial Fibrillation Detection via Parameterised Hypercomplex Neural Networks. in 31st European Signal Processing Conference: EUSIPCO 2023. European Signal Processing Conference, S. 1375-1379, 31st European Signal Processing Conference, EUSIPCO 2023, Helsinki, Finnland, 4 Sept. 2023. https://doi.org/10.48550/arXiv.2211.02678, https://doi.org/10.23919/EUSIPCO58844.2023.10289763
Basso, L., Ren, Z., & Nejdl, W. (2023). Efficient ECG-based Atrial Fibrillation Detection via Parameterised Hypercomplex Neural Networks. In 31st European Signal Processing Conference: EUSIPCO 2023 (S. 1375-1379). (European Signal Processing Conference). https://doi.org/10.48550/arXiv.2211.02678, https://doi.org/10.23919/EUSIPCO58844.2023.10289763
Basso L, Ren Z, Nejdl W. Efficient ECG-based Atrial Fibrillation Detection via Parameterised Hypercomplex Neural Networks. in 31st European Signal Processing Conference: EUSIPCO 2023. 2023. S. 1375-1379. (European Signal Processing Conference). doi: 10.48550/arXiv.2211.02678, 10.23919/EUSIPCO58844.2023.10289763
Basso, Leonie ; Ren, Zhao ; Nejdl, Wolfgang. / Efficient ECG-based Atrial Fibrillation Detection via Parameterised Hypercomplex Neural Networks. 31st European Signal Processing Conference: EUSIPCO 2023. 2023. S. 1375-1379 (European Signal Processing Conference).
Download
@inproceedings{3eb059a33827448fa68718edac035ffb,
title = "Efficient ECG-based Atrial Fibrillation Detection via Parameterised Hypercomplex Neural Networks",
abstract = "Atrial fibrillation (AF) is the most common cardiac arrhythmia and associated with a high risk for serious conditions like stroke. The use of wearable devices embedded with automatic and timely AF assessment from electrocardiograms (ECGs) has shown to be promising in preventing life-threatening situations. Although deep neural networks have demonstrated superiority in model performance, their use on wearable devices is limited by the trade-off between model performance and complexity. In this work, we propose to use lightweight convolutional neural networks (CNNs) with parameterised hypercomplex (PH) layers for AF detection based on ECGs. The proposed approach trains small-scale CNNs, thus overcoming the limited computing resources on wearable devices. We show comparable performance to corresponding real-valued CNNs on two publicly available ECG datasets using significantly fewer model parameters. PH models are more flexible than other hypercomplex neural networks and can operate on any number of input ECG leads.",
keywords = "Atrial fibrillation, ECG analysis, hypercomplex domain, lightweight neural networks",
author = "Leonie Basso and Zhao Ren and Wolfgang Nejdl",
note = "Funding Information: This research was funded by the Federal Ministry of Education and Research (BMBF), Germany under the project LeibnizKILabor with grant No. 01DD20003. ; 31st European Signal Processing Conference, EUSIPCO 2023 ; Conference date: 04-09-2023 Through 08-09-2023",
year = "2023",
doi = "10.48550/arXiv.2211.02678",
language = "English",
isbn = "979-8-3503-2811-0",
series = "European Signal Processing Conference",
pages = "1375--1379",
booktitle = "31st European Signal Processing Conference",

}

Download

TY - GEN

T1 - Efficient ECG-based Atrial Fibrillation Detection via Parameterised Hypercomplex Neural Networks

AU - Basso, Leonie

AU - Ren, Zhao

AU - Nejdl, Wolfgang

N1 - Funding Information: This research was funded by the Federal Ministry of Education and Research (BMBF), Germany under the project LeibnizKILabor with grant No. 01DD20003.

PY - 2023

Y1 - 2023

N2 - Atrial fibrillation (AF) is the most common cardiac arrhythmia and associated with a high risk for serious conditions like stroke. The use of wearable devices embedded with automatic and timely AF assessment from electrocardiograms (ECGs) has shown to be promising in preventing life-threatening situations. Although deep neural networks have demonstrated superiority in model performance, their use on wearable devices is limited by the trade-off between model performance and complexity. In this work, we propose to use lightweight convolutional neural networks (CNNs) with parameterised hypercomplex (PH) layers for AF detection based on ECGs. The proposed approach trains small-scale CNNs, thus overcoming the limited computing resources on wearable devices. We show comparable performance to corresponding real-valued CNNs on two publicly available ECG datasets using significantly fewer model parameters. PH models are more flexible than other hypercomplex neural networks and can operate on any number of input ECG leads.

AB - Atrial fibrillation (AF) is the most common cardiac arrhythmia and associated with a high risk for serious conditions like stroke. The use of wearable devices embedded with automatic and timely AF assessment from electrocardiograms (ECGs) has shown to be promising in preventing life-threatening situations. Although deep neural networks have demonstrated superiority in model performance, their use on wearable devices is limited by the trade-off between model performance and complexity. In this work, we propose to use lightweight convolutional neural networks (CNNs) with parameterised hypercomplex (PH) layers for AF detection based on ECGs. The proposed approach trains small-scale CNNs, thus overcoming the limited computing resources on wearable devices. We show comparable performance to corresponding real-valued CNNs on two publicly available ECG datasets using significantly fewer model parameters. PH models are more flexible than other hypercomplex neural networks and can operate on any number of input ECG leads.

KW - Atrial fibrillation

KW - ECG analysis

KW - hypercomplex domain

KW - lightweight neural networks

UR - http://www.scopus.com/inward/record.url?scp=85178373424&partnerID=8YFLogxK

U2 - 10.48550/arXiv.2211.02678

DO - 10.48550/arXiv.2211.02678

M3 - Conference contribution

AN - SCOPUS:85178373424

SN - 979-8-3503-2811-0

T3 - European Signal Processing Conference

SP - 1375

EP - 1379

BT - 31st European Signal Processing Conference

T2 - 31st European Signal Processing Conference, EUSIPCO 2023

Y2 - 4 September 2023 through 8 September 2023

ER -

Von denselben Autoren