EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme

Mehmet Yiǧit Balik; Kaan Gökçe; Sezgin Atmaca; Emre Aslanger; Arda Güler; Ilkay Öksüz

doi:10.1109/SIU59756.2023.10223856

EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme

Translated title of the contribution: Interpretable Deep Learning for Myocardial Infarction Detection from ECG Signals

Mehmet Yiǧit Balik
, Kaan Gökçe
, Sezgin Atmaca
, Emre Aslanger
, Arda Güler
, Ilkay Öksüz

Department of Computer Engineering

Istanbul Technical University
Mehmet Akif Ersoy Gogus Kalp Ve Damar Cerrahisi Egitim Ve Arastirma Hastanesi
Kardiyoloji Anabilim Dali

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Citations (Scopus)

Abstract

Acute heart attacks such as myocardial infarction (MI) are the main reasons for global deaths. Additionally, approximately half of the deaths occur before the treatment. Hence, it is crucial to diagnose MI fast and cheaply. 12-lead electrocardiogram (ECG) is noninvasive and fast compared to alternative devices. In this work, we aimed to train and validate a residual network model that can distinguish MI and healthy 12-lead ECG records. Moreover, we investigated the contribution of patient information such as age and sex to the decision. Additionally, we compared the performances of models trained with two different loss functions which are binary cross-entropy and pinball loss. We observed the highest accuracy, recall, and F1 score which are 97.86%, 98.73%, and 98.66%, respectively. Furthermore, since we used a convolutional neural network-based architecture, we obtained explainable results using gradient class activation maps by highlighting the ECG segments that contribute the most to the decision.

Translated title of the contribution	Interpretable Deep Learning for Myocardial Infarction Detection from ECG Signals
Original language	Turkish
Title of host publication	31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798350343557
DOIs	https://doi.org/10.1109/SIU59756.2023.10223856
Publication status	Published - 2023
Event	31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023 - Istanbul, Turkey Duration: 5 Jul 2023 → 8 Jul 2023

Publication series

Name	31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023

Conference

Conference	31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023
Country/Territory	Turkey
City	Istanbul
Period	5/07/23 → 8/07/23

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

UN SDGs

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

SDG 3 Good Health and Well-being

Access to Document

10.1109/SIU59756.2023.10223856

Cite this

Balik, M. Y., Gökçe, K., Atmaca, S., Aslanger, E., Güler, A., & Öksüz, I. (2023). EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme. In 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023 (31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SIU59756.2023.10223856

Balik, Mehmet Yiǧit ; Gökçe, Kaan ; Atmaca, Sezgin et al. / EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme. 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023. Institute of Electrical and Electronics Engineers Inc., 2023. (31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023).

@inproceedings{6a0169481fcc4b2a98ed5a728500d43b,

title = "EKG Sinyallerinden Miyokard Enfarkt{\"u}s Tespiti i{\c c}in Yorumlanabilir Derin {\"O}ǧrenme",

abstract = "Acute heart attacks such as myocardial infarction (MI) are the main reasons for global deaths. Additionally, approximately half of the deaths occur before the treatment. Hence, it is crucial to diagnose MI fast and cheaply. 12-lead electrocardiogram (ECG) is noninvasive and fast compared to alternative devices. In this work, we aimed to train and validate a residual network model that can distinguish MI and healthy 12-lead ECG records. Moreover, we investigated the contribution of patient information such as age and sex to the decision. Additionally, we compared the performances of models trained with two different loss functions which are binary cross-entropy and pinball loss. We observed the highest accuracy, recall, and F1 score which are 97.86\%, 98.73\%, and 98.66\%, respectively. Furthermore, since we used a convolutional neural network-based architecture, we obtained explainable results using gradient class activation maps by highlighting the ECG segments that contribute the most to the decision.",

keywords = "Deep Learning, Electrocardiogram, Gradient Class Activation Map, Myocardial Infarction, Residual Network",

author = "Balik, \{Mehmet Yiǧit\} and Kaan G{\"o}k{\c c}e and Sezgin Atmaca and Emre Aslanger and Arda G{\"u}ler and Ilkay {\"O}ks{\"u}z",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023 ; Conference date: 05-07-2023 Through 08-07-2023",

year = "2023",

doi = "10.1109/SIU59756.2023.10223856",

language = "T{\"u}rk{\c c}e",

series = "31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023",

address = "United States",

}

Balik, MY, Gökçe, K, Atmaca, S, Aslanger, E, Güler, A & Öksüz, I 2023, EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme. in 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023. 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023, Institute of Electrical and Electronics Engineers Inc., 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023, Istanbul, Turkey, 5/07/23. https://doi.org/10.1109/SIU59756.2023.10223856

EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme. / Balik, Mehmet Yiǧit; Gökçe, Kaan; Atmaca, Sezgin et al.
31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023. Institute of Electrical and Electronics Engineers Inc., 2023. (31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme

AU - Balik, Mehmet Yiǧit

AU - Gökçe, Kaan

AU - Atmaca, Sezgin

AU - Aslanger, Emre

AU - Güler, Arda

AU - Öksüz, Ilkay

PY - 2023

Y1 - 2023

N2 - Acute heart attacks such as myocardial infarction (MI) are the main reasons for global deaths. Additionally, approximately half of the deaths occur before the treatment. Hence, it is crucial to diagnose MI fast and cheaply. 12-lead electrocardiogram (ECG) is noninvasive and fast compared to alternative devices. In this work, we aimed to train and validate a residual network model that can distinguish MI and healthy 12-lead ECG records. Moreover, we investigated the contribution of patient information such as age and sex to the decision. Additionally, we compared the performances of models trained with two different loss functions which are binary cross-entropy and pinball loss. We observed the highest accuracy, recall, and F1 score which are 97.86%, 98.73%, and 98.66%, respectively. Furthermore, since we used a convolutional neural network-based architecture, we obtained explainable results using gradient class activation maps by highlighting the ECG segments that contribute the most to the decision.

AB - Acute heart attacks such as myocardial infarction (MI) are the main reasons for global deaths. Additionally, approximately half of the deaths occur before the treatment. Hence, it is crucial to diagnose MI fast and cheaply. 12-lead electrocardiogram (ECG) is noninvasive and fast compared to alternative devices. In this work, we aimed to train and validate a residual network model that can distinguish MI and healthy 12-lead ECG records. Moreover, we investigated the contribution of patient information such as age and sex to the decision. Additionally, we compared the performances of models trained with two different loss functions which are binary cross-entropy and pinball loss. We observed the highest accuracy, recall, and F1 score which are 97.86%, 98.73%, and 98.66%, respectively. Furthermore, since we used a convolutional neural network-based architecture, we obtained explainable results using gradient class activation maps by highlighting the ECG segments that contribute the most to the decision.

KW - Deep Learning

KW - Electrocardiogram

KW - Gradient Class Activation Map

KW - Myocardial Infarction

KW - Residual Network

UR - https://www.scopus.com/pages/publications/85173538309

U2 - 10.1109/SIU59756.2023.10223856

DO - 10.1109/SIU59756.2023.10223856

M3 - Konferans katkısı

AN - SCOPUS:85173538309

T3 - 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023

BT - 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023

Y2 - 5 July 2023 through 8 July 2023

ER -

Balik MY, Gökçe K, Atmaca S, Aslanger E, Güler A, Öksüz I. EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme. In 31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023. Institute of Electrical and Electronics Engineers Inc. 2023. (31st IEEE Conference on Signal Processing and Communications Applications, SIU 2023). doi: 10.1109/SIU59756.2023.10223856

EKG Sinyallerinden Miyokard Enfarktüs Tespiti için Yorumlanabilir Derin Öǧrenme

Abstract

Publication series

Conference

Bibliographical note

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this