Automated Classification of Atrioventricular Block Based on Semantic Segmentation of ECG Signals.

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• Atrioventricular block (AVB) is a cardiac condition affecting heart rhythm.
• Accurate detection of P-waves is crucial for diagnosing AVB.
• Current diagnostic methods may require complex equipment or specialized interpretation.

Purpose of the Study:

• To develop a novel algorithm for P-wave detection and AVB classification.
• To utilize a single-channel wearable electrocardiogram (ECG) for accessible cardiac monitoring.
• To enhance the accuracy and ease of diagnosing atrioventricular block.

Main Methods:

• A semantic segmentation algorithm was applied to single-channel wearable ECG data.
• The algorithm was trained and validated for P-wave detection.
• Classification of atrioventricular block was performed based on detected P-wave characteristics.

Main Results:

• The developed algorithm demonstrated accurate P-wave detection capabilities.
• The system successfully classified different types of atrioventricular block.
• The approach is suitable for easily attachable, single-channel wearable ECG devices.

Conclusions:

• A novel semantic segmentation algorithm enables accurate P-wave detection and AVB classification from wearable ECGs.
• This technology offers a promising, non-invasive approach for remote cardiac monitoring and diagnosis.
• The findings support the potential of wearable ECGs for widespread cardiovascular health screening.