ECG Heartbeat Classification Based on an Improved ResNet-18 Model.

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Signal Processing

Background:

• Accurate classification of electrocardiogram (ECG) signals is crucial for diagnosing cardiac conditions.
• Existing deep learning models face challenges in achieving high performance for complex arrhythmia detection.
• Convolutional Neural Networks (CNNs) offer potential for analyzing intricate ECG patterns.

Purpose of the Study:

• To propose an improved ResNet-18 model for enhanced ECG heartbeat classification.
• To leverage the deep layered structure of CNNs with residual connections for improved performance.
• To evaluate the model's effectiveness on a standard arrhythmia database.

Main Methods:

• Development of an improved ResNet-18 model based on a CNN architecture.
• Careful model training and parameter adjustment for optimal performance.
• Application and evaluation of the model on the MIT-BIH arrhythmia database.

Main Results:

• The proposed ResNet-18 model achieved a high overall accuracy of 96.50% on the MIT-BIH dataset.
• Demonstrated superior performance compared to other state-of-the-art classification models.
• Achieved high sensitivity (93.83%) and precision (97.44%) for the ventricular ectopic heartbeat class.

Conclusions:

• The improved ResNet-18 model offers a robust and accurate solution for ECG heartbeat classification.
• The residual structure enables deeper CNN architectures, leading to better diagnostic performance.
• This approach shows significant promise for clinical application in arrhythmia detection.