Atrial Fibrillation Detection with Single-Lead Electrocardiogram Based on Temporal Convolutional Network-ResNet.

Area of Science:

• Cardiology
• Biomedical Engineering
• Artificial Intelligence in Medicine

Background:

• Atrial fibrillation (AFib) is a prevalent cardiac arrhythmia characterized by rapid, irregular atrial rhythms.
• Accurate and timely diagnosis of AFib is crucial for effective patient management and prevention of complications.
• Current diagnostic methods can be enhanced by leveraging advanced computational techniques.

Purpose of the Study:

• To develop and evaluate a novel deep learning model for classifying atrial fibrillation using single-lead electrocardiograms (ECGs).
• To integrate Temporal Convolutional Network (TCN) and Residual Network (ResNet) frameworks for robust AFib detection.
• To improve the accuracy and efficiency of automated AFib classification in clinical settings.

Main Methods:

• Implementation of a hybrid deep learning architecture combining TCN and ResNet.
• Training and validation of the model on single-lead ECG datasets.
• Performance evaluation using key metrics such as accuracy and F1 score.

Main Results:

• The developed model achieved a high accuracy rate of 97% in detecting atrial fibrillation.
• An F1 score of 87% was obtained, demonstrating excellent performance across different classes.
• The results indicate a balanced and accurate classification capability, even for minority classes.

Conclusions:

• The TCN-ResNet hybrid model shows significant promise for the accurate detection of atrial fibrillation from ECGs.
• This advanced neural network approach offers a valuable tool for enhancing cardiac arrhythmia diagnosis.
• The study contributes novel perspectives and technological advancements to the field of cardiology and AI in medicine.