Kardiyolojide Doğru Aritmi Sınıflandırması için Derin Öğrenme Modelleri

Abstract

Kardiyak aritmiler, dünya genelinde kardiyovasküler hastalıklara bağlı mortalitenin önemli bir nedeni olmaya devam etmektedir. Elektrokardiyogram (EKG) verilerinin aritmi sınıflandırması için manuel yorumlanması, doğası gereği öznel ve zaman alıcıdır, bu da doğru ve etkin tanı için zorluklar oluşturur. Derin öğrenmedeki son gelişmeler, EKG analizini otomatikleştirme ve tanısal hassasiyeti artırma konusunda büyük umut vaat etmektedir. Bu çalışma, yaygın olarak kullanılan MIT-BIH Aritmi verisetini kullanarak çeşitli derin öğrenme mimarilerinin performansını değerlendirmektedir. Bulgularımız, ConvLSTM modelinin literatürdeki benzerlerinden çok daha başarılı olarak test setinde %98,81’lik üstün bir doğruluk elde ettiğini göstermektedir. Ayrıca, CNN modeli normal kalp atışlarını etkili bir şekilde tanımlarken, ConvLSTM modeli erken ventriküler kontraksiyonların tespitinde en yüksek performansı göstermektedir. Karmaşık veri ön işleme süreçlerini basitleştirerek, kapsamlı manuel özellik mühendisliği ihtiyacını ortadan kaldırarak otomatik özellik çıkarma imkânı sağlayan derin öğrenme modelleri, aritmi tespitinde önemli bir avantaj sağlamıştır. Bu çalışma, derin öğrenmenin yaygın klinik uygulamalar için potansiyelini vurgulamakta ve hibrit derin öğrenme algoritmalarının gelecekteki tanısal sistemlerde, probleme özgü yüksek performans elde edebileceğini vurgulamaktadır.

Keywords

• Kardiyak aritmi tespiti
• Derin öğrenme
• Tanısal doğruluk
• CNN
• ConvLSTM

Deep Learning Models for Accurate Arrhythmia Classification in Cardiology

Abstract

Cardiac arrhythmias remain a significant contributor to global cardiovascular disease-related mortality. The manual interpretation of electrocardiogram (ECG) data for arrhythmia classification is inherently subjective and time-intensive, posing challenges for accurate and efficient diagnosis. Recent advancements in deep learning have shown great promise in automating ECG analysis and improving diagnostic precision. This study evaluates the performance of various deep learning architectures using the widely recognized MIT-BIH Arrhythmia Dataset. Our findings demonstrate that the ConvLSTM model achieves superior accuracy, reaching 98.81% on the test set. Moreover, while the CNN model effectively identifies normal heartbeats, the ConvLSTM model exhibits the highest performance in detecting premature ventricular contractions. By simplifying complex data preprocessing, eliminating the need for extensive manual feature engineering, and enabling automatic feature extraction, deep learning models offer a transformative approach to arrhythmia detection. This study highlights the potential of deep learning for widespread clinical implementation and suggests that hybrid deep learning algorithms could achieve high problem-specific performance in future diagnostic systems.

Keywords

• Cardiac arrhythmia detection
• CNN
• ConvLSTM
• Deep learning
• Diagnostic accuracy

References

1. Alsayat A., Mahmoud AA., Alanazi S., Mostafa AM., Alshammari N., Alrowaily MA., Shabana H., Ezz M. Enhancing cardiac diagnostics: A deep learning ensemble approach for precise ECG image classification. Journal of Big Data 2025; 12: 7.
2. Bai X., Dong X., Li Y., Liu R. A hybrid deep learning network for automatic diagnosis of cardiac arrhythmia based on 12-lead ECG. Scientific Reports, 2024; 14: Article 24441.
3. Boutellaa E., Kerdjidj O., Ghanem K. Covariance matrix-based fall detection from multiple wearable sensors. Journal of Biomedical Informatics 2019; 94: 103179.
4. Bravo J. Forecasting longevity for financial applications: A first experiment with deep learning methods. Proceedings of the International Conference on Artificial Intelligence and Big Data Analytics for Financial Applications 2021; 1525: 232-249.
5. Burlingame J., Horiuchi B., Ohana P., Onaka A., Sauvage LM. Contribution of heart disease to pregnancy-related deaths based on pregnancy mortality surveillance system. Journal of Perinatology 2012; 32(3): 163-169.
6. Casilari E., Álvarez-Marco M., García-Lagos F. A study of the use of gyroscope measurements in wearable fall detection systems. Symmetry 2020; 12: 649.
7. Cho Y., Kwon JM., Kim KH., Medina-Inojosa JR., Jeon KH., Cho S., Lee SY., Park J., Oh BH. Artificial intelligence algorithm for detecting myocardial infarction using six-lead electrocardiography. Scientific Reports, 2020; 10(1): 20495.
8. Dessein P., Gonzalez-Gay MA. Management of cardiovascular disease risk in rheumatoid arthritis. Journal of Clinical Medicine 2022; 11(12): 3487.

9. Faes MC., Reelick MF., Joosten-Weyn Banningh LW., de Gier M., Esselink RA., Olde Rikkert MG. Qualitative study on the impact of falling in frail older persons and family caregivers: Foundations for an intervention to prevent falls. Aging and Mental Health 2010; 14(7): 834-842.
10. Gustafsson S., Gedon D., Lampa E., Ribeiro AH., Holzmann MJ., Schön TB., Sundström J. Development and validation of deep learning ECG-based prediction of myocardial infarction in emergency department patients. Scientific Reports 2022; 12(1): 24254.
11. Hua Y., Zhao Z., Li R., Chen X., Liu Z., Zhang H. Deep learning with long short-term memory for time series prediction. IEEE Communications Magazine 2019; 57(6): 114-119.
12. Hwang J., Kang J., Jang Y., Kim HC. Development of a novel algorithm and real-time monitoring ambulatory system using a Bluetooth module for fall detection in the elderly. Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and HONGY Biology Society (EMBC) 2004; 3: 2204-2207.
13. Izci E., Ozdemir MA., Degirmenci M., Akan A. Cardiac arrhythmia detection from 2D ECG images by using deep learning technique. Proceedings of the Medical Technologies Congress (TIPTEKNO) 2019; 1-4.
14. Khurshid S., Friedman S., Pirruccello JP., Di Achille P., Diamant N., Anderson CD., Ellinor PT., Batra P., Ho JE., Philippakis AA., Lubitz SA. Deep learning for predicting left ventricular mass and hypertrophy derived from cardiac magnetic resonance using 12-lead ECGs. Circulation: Cardiovascular Imaging 2021; 14(6): 12281.
15. Kim D., Lee KR., Lim DS., Sohn CB. A novel hybrid CNN-transformer model for arrhythmia detection without R-peak identification using stockwell transform. Scientific Reports, 2025; 15: Article 7817.
16. Kostopoulos P., Kyritsis A., Deriaz M., Konstantas D. F2D: A location-aware fall detection system tested with real data from the daily life of elderly people. Procedia Computer Science 2016; 98: 212-219.
17. Kuzuya M., Enoki H., Hasegawa J., Izawa S., Hirakawa Y., Uemura K. Falls of the elderly are associated with the burden of caregivers in the community. International Journal of Geriatric Psychiatry 2006; 21(8): 740-745.
18. Li K., Kaaks R., Linseisen J., Rohrmann S. Associations of dietary calcium intake and calcium supplementation with myocardial infarction, stroke risk, and overall cardiovascular mortality in the EPIC-Heidelberg cohort study. Heart 2012; 98(12): 920-925.
19. Li Q., Stankovic JA., Hanson MA., Barth AT., Lach J., Zhou G. Accurate, fast fall detection using gyroscopes and accelerometer-derived posture information. Proceedings of the 6th International Workshop on Wearable and Implantable Body Sensor Networks (BSN) 2009; 138-143.
20. Lin CS., Lin C., Fang WH., Hsu CJ., Chen SJ., Huang KH., Lin WS., Tsai CS., Kuo CC., Chau T., Yang SJ., Lin SH. A deep learning algorithm (EKG12Net) for detecting hypokalemia and hyperkalemia via electrocardiography: Algorithm development. JMIR Medical Informatics 2020; 8(3): e15931.
21. Mattiuzzi C., Lippi G. Gender-based fatal effects of environmental air pollution. Environmental Science and Pollution Research 2020; 27(10): 11458.
22. Midani W., Ouarda W., Ltifi H., Ben Ayed, M. S2SDeepArr: Sequence to sequence deep learning architecture for arrhythmia detection under the inter-patient paradigm. Procedia Computer Science 2024; 246: 792–801.
23. Moody GB., Mark RG. The impact of the MIT-BIH Arrhythmia Database. IEEE Engineering in Medicine and Biology Magazine 2001; 20(3): 45-50.
24. Mujib M., Zhang Y., Feller MA., Ahmed A. Evidence of a “Heart Failure Belt” in the Southeastern United States. American Journal of Cardiology 2011; 107(6): 935-937.
25. Obeng-Gyasi E., Ferguson AC., Stamatakis KA., Province MA. The combined effect of lead exposure and allostatic load on cardiovascular disease mortality: A pilot study. International Journal of Environmental Research and Public Health 2021; 18(13): 6879.
26. Ojha MK., Wadhwani S., Wadhwani AK., Shukla A. Automated detection of arrhythmias from an ECG signal using an autoencoder and SVM classifier. Physical and Engineering Sciences in Medicine 2022; 45(2): 665-674.
27. Qiu C., Johansson G., Zhu F., Kivipelto M., Winblad B. Prevention of cognitive decline in old age:Varying effects of interventions in different populations. Annals of Translational Medicine 2019; 7(Suppl 3): S3.
28. Quadros T., Lazzaretti A., Schneider F. A movement decomposition and machine learning-based fall detection system using a wrist wearable device. IEEE Sensors Journal 2018; 18(22): 1-1.
29. Raj S., Ray KC. A personalized arrhythmia monitoring platform. Scientific Reports 2018; 8: 11395.
30. Rajkumar A., Ganesan M., Lavanya R. Arrhythmia classification on ECG using deep learning. Proceedings of the 5th International Conference on Advanced Computing and Communication Systems (ICACCS) 2019; 365-369.
31. Roglic G., Unwin N., Bennett PH., Mathers C., Tuomilehto J., Nag S., Connolly V., King H. The burden of mortality attributable to diabetes: Realistic estimates for the year 2000. Diabetes Care 2005; 28(9): 2130-2135.
32. Salah O., Abdulla R., Selvaperumal SK. Accelerometer-based elderly fall detection system using edge artificial intelligence architecture. International Journal of Electrical and Computer Engineering 2022; 12: 4430-4438.
33. Sawano S., Kodera S., Katsushika S., Nakamoto M., Ninomiya K., Shinohara H., Higashikuni Y., Nakanishi K., Nakao T., Seki T., Takeda N., Fujiu K., Daimon M., Akazawa H., Morita H., Komuro I. Deep learning model for detecting significant aortic regurgitation using electrocardiography. Journal of Cardiology 2022; 79(3): 334-341.
34. Shahzad A., Kim K. FallDroid: An automated smartphone-based fall detection system using multiple kernel learning. IEEE Transactions on Industrial Informatics 2018; 15(1): 35-44.
35. Shakeri S. A smartphone-based fall detection system using accelerometer and microphone. Semantic Scholar 2017.
36. Silva DAS., Malta DC., Souza MFM., Naghavi M. Burden of ischemic heart disease mortality attributable to physical inactivity in Brazil. Revista de Saúde Pública 2018; 52: 72.
37. Suriani NS., Rashid FN., Yunos NY. Optimal accelerometer placement for fall detection of rehabilitation patients. Journal of Telecommunication, Electronic and Computer Engineering 2018; 10(2-5): 25-30.
38. Vaswani A., Shazeer N., Parmar N., Uszkoreit J., Jones L., Gomez AN., Kaiser Ł., Polosukhin I. Attention is all you need. Advances in Neural Information Processing Systems 2017; 30: 1-11.
39. Wang H., Zhang D., Wang Y., Ma J., Wang Y., Li S. RT-Fall: A real-time and contactless fall detection system with commodity Wi-Fi devices. IEEE Transactions on Mobile Computing 2016; 16: 1-1.
40. Yacchirema D., Puga J., Palau C., Esteve M. Fall detection system for elderly people using IoT and Big Data. Procedia Computer Science 2018; 130: 603-610.
41. Yacoub S., Kotit S., Yacoub MH. The outlook and evolution of disease against a background of climate change and dwindling resources. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 2011; 369(1942): 1719-1729.
42. Yoon T., Kang D. Multimodal stacking ensemble for cardiovascular disease diagnosis. Journal of Personalized Medicine 2023; 13(2): 373.
43. Yu X. Approaches and principles of fall detection for elderly and patient. Proceedings of the IEEE International Conference on Health 2008; 47: 42-47.
44. Yu X., Zeng N., Liu S., Zhang YD. Utilization of DenseNet201 for the diagnosis of breast abnormality. Machine Vision and Applications 2019; 30(7): 1135-1144.
45. Yusuf M., Farooq H., Wood K., Fidler G., Yang C., Zou X. Human sensing in reverberant environments: RF-based occupancy and fall detection in ships. IEEE Transactions on Vehicular Technology 2021; 70: 1-1.
46. Zhang Q., Ren L., Shi W. HONEY: A multimodality fall detection and telecare system. Telemedicine and e-Health 2013; 19: 415-422.
47. Zhu H., Samtani S., Brown R. A deep learning approach for recognizing activity of daily living (ADL) for senior care: Exploiting interaction dependency and temporal patterns. Management Information Systems Quarterly 2021; 45: 859-896.