Performance Analysis of Ten Common QRS Detectors on Different ECG Application Cases.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Accurate QRS complex detection is crucial for electrocardiogram (ECG) analysis.
• Numerous QRS detection algorithms exist, but their performance varies across different clinical scenarios.
• A systematic evaluation is needed to understand algorithm utility in diverse applications.

Purpose of the Study:

• To systematically evaluate the performance and utility of ten widely used QRS detection algorithms.
• To compare algorithm effectiveness across varying ECG signal qualities (high vs. low).
• To assess algorithm robustness against arrhythmias, paced rhythms, and dynamic telehealth signals.

Main Methods:

• Conducted four experiments using six internationally recognized ECG databases.
• Evaluated algorithm accuracy (F1 score) and computational complexity (time cost).
• Tested algorithms on high-quality, low-quality, normal, arrhythmic, paced rhythm, and telehealth ECG datasets.

Main Results:

• All algorithms performed well (>99% F1) on high-quality ECGs, but accuracy dropped significantly (<80% F1) on low-quality signals.
• Algorithms generally showed good performance (>95% F1) on normal and arrhythmic ECGs, with minor exceptions.
• Most algorithms were robust to paced beats (>94% F1), except the RS slope method (78.99% F1).
• Telehealth ECG signals posed challenges, with most algorithms <80% F1, except OKB (80.43% F1).
• Computational efficiency was high for most algorithms (<4ms), with OKB being the fastest (1.54ms).

Conclusions:

• QRS detection algorithm performance is highly dependent on ECG signal quality and application context.
• The OKB algorithm demonstrates superior performance and efficiency, particularly in challenging telehealth scenarios.
• Algorithm selection should consider signal quality, presence of arrhythmias or pacing, and real-time processing requirements.