QRS detection based on wavelet coefficients.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Electrocardiogram (ECG) signal analysis is vital for assessing heart function.
• Accurate QRS complex detection is fundamental for automated ECG feature extraction.
• Existing QRS detectors face challenges with diverse QRS morphologies.

Purpose of the Study:

• To investigate the efficacy of wavelet detail coefficients for detecting various QRS complex morphologies in ECG signals.
• To develop a robust QRS detection method capable of differentiating true (normal and abnormal) beats from false beats.

Main Methods:

• Utilized wavelet detail coefficients for ECG signal analysis.
• Employed the power spectrum of QRS complexes across different energy levels to identify beat characteristics.
• Developed a discrimination strategy based on power spectrum differences between normal, abnormal, and false beats.

Main Results:

• The proposed method demonstrated significant performance enhancement on the MIT-BIH Arrhythmia Database (MITDB).
• Achieved a sensitivity of 99.64% for QRS complex detection.
• Obtained a positive predictivity of 99.82% in distinguishing true from false beats.

Conclusions:

• Wavelet detail coefficients offer a powerful tool for accurate QRS complex detection in ECG.
• The proposed power spectrum-based method effectively discriminates diverse QRS morphologies.
• This approach significantly improves the reliability of automated ECG analysis for arrhythmia detection.