Key result
Deep learning neural networks improve accuracy and reliability in automatic ECG arrhythmia classification.
Why the study?
Recent advances in classification algorithms and deep learning have significantly improved automatic ECG arrhythmia classification, warranting a systematic analysis of neural network and non-neural network models under standardized criteria.
This review highlights that neural network-based deep learning technologies significantly enhance the accuracy and reliability of automatic ECG arrhythmia classification.
Supports DL integration in ECG monitoring; leaves open need for prospective clinical validation.
ECG (Electrocardiogram), the most commonly used tool in the diagnosis of cardiac diseases, contains a large amount of physiologic information about the electrical activity of the heart. Research on automatic diagnosis of cardiac diseases by means of computer-aided ECG diagnosis has been carried out for decades. Computer assisted therapy is able to timely detect heart diseases and reduce the mortality rate of cardiovascular disease patients. In recent years, classification algorithms applied to automatic ECG arrhythmia diagnosis have been proposed and optimized. With the application and development of neural network-based deep learning technology in automatic ECG diagnosis, the accuracy and reliability of automatic ECG arrhythmia classification have been significantly improved. This paper systematically analyzes the literatures related to the automatic classification of arrhythmias based on ECG signals, summarizes the classification algorithms of different arrhythmias based on neural network model and non-neural network model, and summarizes the classification characteristics of the two main classification models in the application of automatic diagnosis of arrhythmias under the same AAMI standard and MIT-BIH arrhythmia database. This paper summarizes the selected research results in the field of arrhythmia classification under the same database from the perspective of data quality, evaluation paradigm and performance indicators, and discusses the correlation between the specific characteristics of arrhythmia detection and different classification algorithms. Finally, the existing problems in the research of automatic classification of arrhythmias are put forward, which provides a reference for the future development and clinical practice of automatic diagnosis of arrhythmias.
No takes yet. Share an insight, caveat, or question.
Sun et al. (2024) conducted a review in ECG Arrhythmias. Classification algorithms (neural network and non-neural network models) was evaluated on Accuracy and reliability of automatic ECG arrhythmia classification. Neural network-based deep learning technology significantly improves the accuracy and reliability of automatic ECG arrhythmia classification.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: