Artificial Intelligence-Enabled ECG Algorithm Based on Improved Residual Network for Wearable ECG

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Cardiology

Background:

• Heart disease remains a leading global cause of mortality.
• Advancements in electrocardiogram (ECG) technology, particularly residual networks (ResNet), have improved cardiac physiology understanding.
• Wearable ECG devices offer continuous monitoring but require sophisticated analysis for accurate diagnosis.

Purpose of the Study:

• To develop an artificial intelligence-enabled ECG algorithm for a wearable device.
• To improve the accuracy and speed of arrhythmia classification using an enhanced ResNet model.
• To integrate a wearable ECG system with a cloud platform for diagnostics.

Main Methods:

• A wearable ECG system comprising conductive fabric electrodes, a wireless acquisition module, a mobile app, and a cloud platform was utilized.
• A novel algorithm based on an improved ResNet-50 architecture was developed for arrhythmia classification.
• ECG signals were converted into two-dimensional images using Gramian angular fields, and the ResNet model was optimized with multistage shortcut branches and SELU activation functions.

Main Results:

• The improved ResNet algorithm achieved an average recognition rate of 98.3% for classifying seven types of arrhythmia.
• The system demonstrated effective rapid classification of various cardiac rhythm abnormalities.
• The integration of the algorithm with the wearable hardware and cloud platform facilitated a comprehensive diagnostic solution.

Conclusions:

• The proposed AI-enabled ECG algorithm based on an improved ResNet significantly enhances the diagnostic capabilities of wearable ECG devices.
• This technology offers a promising approach for early and accurate detection of heart arrhythmias, potentially reducing mortality rates.
• The system provides a scalable solution for remote cardiac monitoring and diagnostics.