QRS complex detection in ECG signals using locally adaptive weighted total variation denoising

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Accurate QRS complex detection is crucial for electrocardiogram (ECG) analysis and clinical diagnosis.
• Challenges in QRS detection include signal noise, artifacts, and interference from other ECG waves (P and T).

Purpose of the Study:

• To propose a novel, adaptive QRS detection technique using weighted total variation (WTV) denoising.
• To improve QRS detection accuracy by simultaneously reducing noise and P- and T-wave interference.

Main Methods:

• ECG preprocessing using locally adaptive weighted total variation (WTV) denoising.
• WTV minimization with a regularization parameter determined by local noise estimation.
• Weighting strategy prioritizing QRS complex preservation during smoothing.

Main Results:

• Achieved high detection accuracy on the MIT-BIH arrhythmia database.
• Offline implementation: 99.90% sensitivity, 99.88% positive predictivity, 0.23% error rate.
• Real-time implementation: 99.86% sensitivity, 99.85% positive predictivity, 0.29% error rate.

Conclusions:

• The proposed WTV denoising technique offers superior QRS detection accuracy compared to existing methods.
• The method exhibits low computational load, suitable for both fast offline and real-time ECG analysis.
• This adaptive denoising approach effectively addresses challenges posed by noise and wave interference in ECG signals.