1. MMC
  2. All Volumes and Issues
  3. Volume 11, Number 3, 2024
  4. AI-Enhanced ECG diagnosis system for acute myocardial infarction with LBBB: Constant-Q transform and ResNet-50 integration

AI-Enhanced ECG diagnosis system for acute myocardial infarction with LBBB: Constant-Q transform and ResNet-50 integration

MMC.
2024;
Volume 11, Number 3
: pp. 654–662
https://doi.org/10.23939/mmc2024.03.654
Received: December 24, 2023
Revised: August 10, 2024
Accepted: August 12, 2024

Elfatouaki H., Adnane L., Charafeddine A. Z., Mohamed A. AI-Enhanced ECG diagnosis system for acute myocardial infarction with LBBB: Constant-Q transform and ResNet-50 integration.  Mathematical Modeling and Computing. Vol. 11, No. 3, pp. 654–662 (2024)

Authors:
  1. H. Elfatouaki
  2. L. Adnane
  3. A. Z. Charafeddine
  4. A. Mohamed
1
National School of Applied Sciences, University of Cadi Ayyad, Marrakesh, Morocco
2
National School of Applied Sciences, University of Cadi Ayyad, Marrakesh, Morocco
3
Polydisciplinary Faculty of Sidi Bennour, Chouaib Doukkali University, Morocco
4
National School of Applied Sciences, University of Cadi Ayyad, Marrakesh, Morocco

This study introduces an advanced Electrocardiogram (ECG) diagnostic framework that melds signal processing techniques with deep learning models to significantly boost accuracy in identifying acute myocardial infarction (MI) and MI related to left bundle branch block (LBBB).  By merging the Constant-Q Transform (CQT) with a pre-trained model, this system showcases exceptional performance, an impressive 98.99% accuracy and a remarkably low 0.0029% training loss after 100 trained epochs.  Rigorous 10-fold cross-validation substantiates and fortifies these findings.  This novel approach streamlines the complexities of diagnostics by consolidating 12-lead ECG data and harnessing CQT for precise time-frequency domain analysis.  Notably, this methodology not only enhances MI detection accuracy but also presents potential for enhancing healthcare outcomes.  It holds promise in minimizing misdiagnoses, thereby propelling advancements in patient care for critical cardiac conditions.  This paradigm shift marks a significant stride in ECG-based diagnostic systems, offering far-reaching implications for improved medical practices and patient well-being.

AMI
LBBB
CQT
ECG
ResNet-50
  1. Ibrahim L., Mesinovic M., Yang K.-W., Eid M. A.  Explainable Prediction of Acute Myocardial Infarction using Machine Learning and Shapley Values.  IEEE Access.  8, 210410–210417 (2020).
  2. Laslett L. J., Alagona P., Clark B. A., Drozda J. P., Saldivar F., Wilson S. R., Poe C., Hart M.  The Worldwide Environment of Cardiovascular Disease: Prevalence, Diagnosis, Therapy, and Policy Issues: A Report From the American College of Cardiology.  Journal of the American College of Cardiology.  60 (25, Supplement), S1–S49 (2012).
  3. Benjamin E. J., Muntner P., Alonso A., Bittencourt M. S., Callaway C. W., et al.  Heart Disease and Stroke Statistics-2019 Update: A Report From the American Heart Association.  Circulation.  139 (10), e56–e528 (2019).
  4. Hollander G., Nadiminti V., Lichstein E., Greengart A., Sanders M.  Bundle branch block in acute myocardial infarction.  American Heart Journal.  105 (5), 738–743 (1983).
  5. Han C., Shi L.  ML-ResNet: A novel network to detect and locate myocardial infarction using 12 leads ECG.  Computer Methods and Programs in Biomedicine.  185, 105138 (2020).
  6. Sakli N., Ghabri H., Soufiene B. O., Almalki F., Sakli H., Ali O., Najjari M.  ResNet-50 for 12-lead electrocardiogram automated diagnosis.  Computational Intelligence and Neuroscience.  2022, 7617551 (2022).
  7. Hao P., Gao X., Li Z., Zhang J., Wu F., Bai C.  Multi-branch Fusion Network for Myocardial Infarction Screening from 12-lead ECG Images.  Computer Methods and Programs in Biomedicine.  184, 105286 (2020).
  8. Hawkins D. M.  The problem of overfitting.  Journal of Chemical Information and Computer Sciences.  44 (1), 1–12 (2004).
  9. Xie S., Girshick R., Dollar P., Tu Z., He K.  Aggregated Residual Transformations for Deep Neural Networks.  2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 5987–5995 (2017).
  10. Berkaya S. K., Uysal A. K., Gunal E. S., Ergin S., Gunal S., Gulmezoglu M. B.  A survey on ECG analysis.  Biomedical Signal Processing and Control.  43, 216–235 (2018).
  11. Youngberg J., Boll S.  Constant-Q signal analysis and synthesis.  ICASSP'78. IEEE International Conference on Acoustics, Speech, and Signal Processing.  375–378 (1978).
  12. Brown J. C.  Calculation of a constant $Q$ spectral transform.  Journal of the Acoustical Society of America.  89 (1), 425–434 (1991).
  13. Barut Z., Altuntaş V.  Comparison of Performance of Different K Values with K-Fold Cross Validation in a Graph-Based Learning Model for lncRNA-Disease Prediction.  Kirklareli Üniversitesi Mühendislik Ve Fen Bilimleri Dergisi.  9 (1), 63–82 (2023).