Atrial fibrillation classification based on convolutional neural networks

Kwang Sig Lee, Sunghoon Jung, Yeongjoon Gil, Ho Sung Son

Research output: Contribution to journalArticle

Abstract

Background: The global age-adjusted mortality rate related to atrial fibrillation (AF) registered a rapid growth in the last four decades, i.e., from 0.8 to 1.6 and 0.9 to 1.7 per 100,000 for men and women during 1990-2010, respectively. In this context, this study uses convolutional neural networks for classifying (diagnosing) AF, employing electrocardiogram data in a general hospital. Methods: Data came from Anam Hospital in Seoul, Korea, with 20,000 unique patients (10,000 normal sinus rhythm and 10,000 AF). 30 convolutional neural networks were applied and compared for the diagnosis of the normal sinus rhythm vs. AF condition: 6 Alex networks with 5 convolutional layers, 3 fully connected layers and the number of kernels changing from 3 to 256; and 24 residual networks with the number of residuals blocks (or kernels) varying from 8 to 2 (or 64 to 2). Results: In terms of the accuracy, the best Alex network was one with 24 initial kernels (i.e., kernels in the first layer), 5,268,818 parameters and the training time of 89 s (0.997), while the best residual network was one with 6 residual blocks, 32 initial kernels, 248,418 parameters and the training time of 253 s (0.999). In general, the performance of the residual network improved as the number of its residual blocks (its depth) increased. Conclusion: For AF diagnosis, the residual network might be a good model with higher accuracy and fewer parameters than its Alex-network counterparts.

Original languageEnglish
Article number206
JournalBMC Medical Informatics and Decision Making
Volume19
Issue number1
DOIs
Publication statusPublished - 2019 Oct 29

Fingerprint

Delayed Emergence from Anesthesia
Atrial Fibrillation
Korea
General Hospitals
Electrocardiography
Mortality
Growth

Keywords

  • Alex networks
  • Atrial fibrillation
  • Convolutional neural networks
  • Residual networks

ASJC Scopus subject areas

  • Health Policy
  • Health Informatics

Cite this

Atrial fibrillation classification based on convolutional neural networks. / Lee, Kwang Sig; Jung, Sunghoon; Gil, Yeongjoon; Son, Ho Sung.

In: BMC Medical Informatics and Decision Making, Vol. 19, No. 1, 206, 29.10.2019.

Research output: Contribution to journalArticle

@article{acdf5dd94dde422695d09a3e098a4223,
title = "Atrial fibrillation classification based on convolutional neural networks",
abstract = "Background: The global age-adjusted mortality rate related to atrial fibrillation (AF) registered a rapid growth in the last four decades, i.e., from 0.8 to 1.6 and 0.9 to 1.7 per 100,000 for men and women during 1990-2010, respectively. In this context, this study uses convolutional neural networks for classifying (diagnosing) AF, employing electrocardiogram data in a general hospital. Methods: Data came from Anam Hospital in Seoul, Korea, with 20,000 unique patients (10,000 normal sinus rhythm and 10,000 AF). 30 convolutional neural networks were applied and compared for the diagnosis of the normal sinus rhythm vs. AF condition: 6 Alex networks with 5 convolutional layers, 3 fully connected layers and the number of kernels changing from 3 to 256; and 24 residual networks with the number of residuals blocks (or kernels) varying from 8 to 2 (or 64 to 2). Results: In terms of the accuracy, the best Alex network was one with 24 initial kernels (i.e., kernels in the first layer), 5,268,818 parameters and the training time of 89 s (0.997), while the best residual network was one with 6 residual blocks, 32 initial kernels, 248,418 parameters and the training time of 253 s (0.999). In general, the performance of the residual network improved as the number of its residual blocks (its depth) increased. Conclusion: For AF diagnosis, the residual network might be a good model with higher accuracy and fewer parameters than its Alex-network counterparts.",
keywords = "Alex networks, Atrial fibrillation, Convolutional neural networks, Residual networks",
author = "Lee, {Kwang Sig} and Sunghoon Jung and Yeongjoon Gil and Son, {Ho Sung}",
year = "2019",
month = "10",
day = "29",
doi = "10.1186/s12911-019-0946-1",
language = "English",
volume = "19",
journal = "BMC Medical Informatics and Decision Making",
issn = "1472-6947",
publisher = "BioMed Central",
number = "1",

}

TY - JOUR

T1 - Atrial fibrillation classification based on convolutional neural networks

AU - Lee, Kwang Sig

AU - Jung, Sunghoon

AU - Gil, Yeongjoon

AU - Son, Ho Sung

PY - 2019/10/29

Y1 - 2019/10/29

N2 - Background: The global age-adjusted mortality rate related to atrial fibrillation (AF) registered a rapid growth in the last four decades, i.e., from 0.8 to 1.6 and 0.9 to 1.7 per 100,000 for men and women during 1990-2010, respectively. In this context, this study uses convolutional neural networks for classifying (diagnosing) AF, employing electrocardiogram data in a general hospital. Methods: Data came from Anam Hospital in Seoul, Korea, with 20,000 unique patients (10,000 normal sinus rhythm and 10,000 AF). 30 convolutional neural networks were applied and compared for the diagnosis of the normal sinus rhythm vs. AF condition: 6 Alex networks with 5 convolutional layers, 3 fully connected layers and the number of kernels changing from 3 to 256; and 24 residual networks with the number of residuals blocks (or kernels) varying from 8 to 2 (or 64 to 2). Results: In terms of the accuracy, the best Alex network was one with 24 initial kernels (i.e., kernels in the first layer), 5,268,818 parameters and the training time of 89 s (0.997), while the best residual network was one with 6 residual blocks, 32 initial kernels, 248,418 parameters and the training time of 253 s (0.999). In general, the performance of the residual network improved as the number of its residual blocks (its depth) increased. Conclusion: For AF diagnosis, the residual network might be a good model with higher accuracy and fewer parameters than its Alex-network counterparts.

AB - Background: The global age-adjusted mortality rate related to atrial fibrillation (AF) registered a rapid growth in the last four decades, i.e., from 0.8 to 1.6 and 0.9 to 1.7 per 100,000 for men and women during 1990-2010, respectively. In this context, this study uses convolutional neural networks for classifying (diagnosing) AF, employing electrocardiogram data in a general hospital. Methods: Data came from Anam Hospital in Seoul, Korea, with 20,000 unique patients (10,000 normal sinus rhythm and 10,000 AF). 30 convolutional neural networks were applied and compared for the diagnosis of the normal sinus rhythm vs. AF condition: 6 Alex networks with 5 convolutional layers, 3 fully connected layers and the number of kernels changing from 3 to 256; and 24 residual networks with the number of residuals blocks (or kernels) varying from 8 to 2 (or 64 to 2). Results: In terms of the accuracy, the best Alex network was one with 24 initial kernels (i.e., kernels in the first layer), 5,268,818 parameters and the training time of 89 s (0.997), while the best residual network was one with 6 residual blocks, 32 initial kernels, 248,418 parameters and the training time of 253 s (0.999). In general, the performance of the residual network improved as the number of its residual blocks (its depth) increased. Conclusion: For AF diagnosis, the residual network might be a good model with higher accuracy and fewer parameters than its Alex-network counterparts.

KW - Alex networks

KW - Atrial fibrillation

KW - Convolutional neural networks

KW - Residual networks

UR - http://www.scopus.com/inward/record.url?scp=85074345558&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85074345558&partnerID=8YFLogxK

U2 - 10.1186/s12911-019-0946-1

DO - 10.1186/s12911-019-0946-1

M3 - Article

C2 - 31664990

AN - SCOPUS:85074345558

VL - 19

JO - BMC Medical Informatics and Decision Making

JF - BMC Medical Informatics and Decision Making

SN - 1472-6947

IS - 1

M1 - 206

ER -