An InceptionTime-Inspired Convolutional Neural Network to Detect Cardiac Abnormalities in Reduced-Lead ECG Data

Harry J. Crocker; Aaron W. Costall

doi:10.23919/CinC53138.2021.9662678

An InceptionTime-Inspired Convolutional Neural Network to Detect Cardiac Abnormalities in Reduced-Lead ECG Data

Harry J. Crocker, Aaron W. Costall

Research output: Chapter or section in a book/report/conference proceeding › Chapter in a published conference proceeding

3 Citations (SciVal)

128 Downloads (Pure)

Abstract

Cardiovascular disease is the leading cause of death worldwide. The twelve-lead electrocardiogram (ECG) is a common tool for diagnosing cardiac abnormalities, but its interpretation requires a trained cardiologist. Thus there is growing interest in automated ECG diagnosis, especially using fewer leads. Hence the PhysioNet-CinC Challenge 2021: Will two (leads) do? The University of Bath team (UoB HBC) developed InceptionTime-inspired deep convolutional neural networks, using parallel 1D convolutions of varying length, for twelve-, six-, four-, three-, and two-lead models. The twelve-lead model achieved a Challenge metric score of 0.35 on the test set, placing the University of Bath team 23rd out of 39 entries. Though the twelve-lead model performed best, three-lead performance was lower by only 0.25 %, suggesting potential for reliable reduced-lead diagnoses. Furthermore, the three-lead model performed consistently better than the six-lead, highlighting the importance of selection of type of lead, not just their number.

Original language	English
Title of host publication	Computing in Cardiology 2021
Publisher	IEEE
Number of pages	4
Volume	48
ISBN (Electronic)	9781665479165
DOIs	https://doi.org/10.23919/CinC53138.2021.9662678
Publication status	E-pub ahead of print - 10 Jan 2022
Event	Computing in Cardiology 2021 - Brno, Czech Republic Duration: 12 Sept 2021 → 15 Sept 2021 http://www.cinc2021.org/

Publication series

Name	Computing in Cardiology
Volume	2021-September
ISSN (Print)	2325-8861
ISSN (Electronic)	2325-887X

Conference

Conference	Computing in Cardiology 2021
Country/Territory	Czech Republic
City	Brno
Period	12/09/21 → 15/09/21
Internet address	http://www.cinc2021.org/

ASJC Scopus subject areas

General Computer Science
Cardiology and Cardiovascular Medicine

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.23919/CinC53138.2021.9662678

CinC_Final_Submission_ID_136_Crocker_Costall.PDF
© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.
Accepted author manuscript, 2.24 MB

Cite this

An InceptionTime-Inspired Convolutional Neural Network to Detect Cardiac Abnormalities in Reduced-Lead ECG Data. / Crocker, Harry J.; Costall, Aaron W.
Computing in Cardiology 2021. Vol. 48 IEEE, 2022. (Computing in Cardiology; Vol. 2021-September).

Research output: Chapter or section in a book/report/conference proceeding › Chapter in a published conference proceeding

@inproceedings{801c2942b313465d8a99d125f804f2ee,

title = "An InceptionTime-Inspired Convolutional Neural Network to Detect Cardiac Abnormalities in Reduced-Lead ECG Data",

abstract = "Cardiovascular disease is the leading cause of death worldwide. The twelve-lead electrocardiogram (ECG) is a common tool for diagnosing cardiac abnormalities, but its interpretation requires a trained cardiologist. Thus there is growing interest in automated ECG diagnosis, especially using fewer leads. Hence the PhysioNet-CinC Challenge 2021: Will two (leads) do? The University of Bath team (UoB HBC) developed InceptionTime-inspired deep convolutional neural networks, using parallel 1D convolutions of varying length, for twelve-, six-, four-, three-, and two-lead models. The twelve-lead model achieved a Challenge metric score of 0.35 on the test set, placing the University of Bath team 23rd out of 39 entries. Though the twelve-lead model performed best, three-lead performance was lower by only 0.25 %, suggesting potential for reliable reduced-lead diagnoses. Furthermore, the three-lead model performed consistently better than the six-lead, highlighting the importance of selection of type of lead, not just their number.",

author = "Crocker, {Harry J.} and Costall, {Aaron W.}",

year = "2022",

month = jan,

day = "10",

doi = "10.23919/CinC53138.2021.9662678",

language = "English",

volume = "48",

series = "Computing in Cardiology",

publisher = "IEEE",

booktitle = "Computing in Cardiology 2021",

address = "USA United States",

note = "Computing in Cardiology 2021 ; Conference date: 12-09-2021 Through 15-09-2021",

url = "http://www.cinc2021.org/",

}

TY - GEN

T1 - An InceptionTime-Inspired Convolutional Neural Network to Detect Cardiac Abnormalities in Reduced-Lead ECG Data

AU - Crocker, Harry J.

AU - Costall, Aaron W.

PY - 2022/1/10

Y1 - 2022/1/10

N2 - Cardiovascular disease is the leading cause of death worldwide. The twelve-lead electrocardiogram (ECG) is a common tool for diagnosing cardiac abnormalities, but its interpretation requires a trained cardiologist. Thus there is growing interest in automated ECG diagnosis, especially using fewer leads. Hence the PhysioNet-CinC Challenge 2021: Will two (leads) do? The University of Bath team (UoB HBC) developed InceptionTime-inspired deep convolutional neural networks, using parallel 1D convolutions of varying length, for twelve-, six-, four-, three-, and two-lead models. The twelve-lead model achieved a Challenge metric score of 0.35 on the test set, placing the University of Bath team 23rd out of 39 entries. Though the twelve-lead model performed best, three-lead performance was lower by only 0.25 %, suggesting potential for reliable reduced-lead diagnoses. Furthermore, the three-lead model performed consistently better than the six-lead, highlighting the importance of selection of type of lead, not just their number.

AB - Cardiovascular disease is the leading cause of death worldwide. The twelve-lead electrocardiogram (ECG) is a common tool for diagnosing cardiac abnormalities, but its interpretation requires a trained cardiologist. Thus there is growing interest in automated ECG diagnosis, especially using fewer leads. Hence the PhysioNet-CinC Challenge 2021: Will two (leads) do? The University of Bath team (UoB HBC) developed InceptionTime-inspired deep convolutional neural networks, using parallel 1D convolutions of varying length, for twelve-, six-, four-, three-, and two-lead models. The twelve-lead model achieved a Challenge metric score of 0.35 on the test set, placing the University of Bath team 23rd out of 39 entries. Though the twelve-lead model performed best, three-lead performance was lower by only 0.25 %, suggesting potential for reliable reduced-lead diagnoses. Furthermore, the three-lead model performed consistently better than the six-lead, highlighting the importance of selection of type of lead, not just their number.

UR - http://www.scopus.com/inward/record.url?scp=85124740152&partnerID=8YFLogxK

U2 - 10.23919/CinC53138.2021.9662678

DO - 10.23919/CinC53138.2021.9662678

M3 - Chapter in a published conference proceeding

AN - SCOPUS:85124740152

VL - 48

T3 - Computing in Cardiology

BT - Computing in Cardiology 2021

PB - IEEE

T2 - Computing in Cardiology 2021

Y2 - 12 September 2021 through 15 September 2021

ER -

An InceptionTime-Inspired Convolutional Neural Network to Detect Cardiac Abnormalities in Reduced-Lead ECG Data

Abstract

Publication series

Conference

ASJC Scopus subject areas

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this