Adaptive learning based heartbeat classification

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Cardiovascular diseases (CVD) are a major cause of mortality and hospitalizations.
• Timely intervention is crucial for reducing CVD morbidity and healthcare costs.
• Manual electrocardiogram (ECG) analysis is time-consuming and labor-intensive.

Purpose of the Study:

• To develop an automated system for detecting abnormal heartbeats from ECG signals.
• To introduce novel time and frequency domain features for improved heartbeat classification.
• To reduce variations in ECG signals for more accurate automated analysis.

Main Methods:

• Extraction of new features from time and frequency domains of ECG signals.
• Application of feature normalization techniques to minimize inter- and intra-patient variability.
• Utilizing an adaptive learning-based classifier for heartbeat detection.

Main Results:

• The proposed method achieves performance comparable to existing literature.
• In many cases, the new approach demonstrates improved accuracy in heartbeat detection.
• The system successfully eliminates the requirement for manual signal labeling by experts.

Conclusions:

• Novel ECG features and normalization techniques enhance automated abnormal heartbeat detection.
• The developed system offers a more efficient and accurate alternative to manual ECG analysis.
• This approach supports timely cardiovascular disease intervention and potentially reduces healthcare burdens.