Exploratory data analysis based efficient QRS-complex detection technique with minimal computational load.

Area of Science:

• Biomedical Engineering
• Cardiology
• Signal Processing

Background:

• Accurate QRS-complex detection in electrocardiograms (ECG) is crucial for diagnosing cardiac disorders.
• Challenges include powerline interference, baseline drift, and variable peak amplitudes.
• Existing methods often require complex parameter tuning and training.

Purpose of the Study:

• To develop an efficient QRS-complex detection technique with minimal computational load.
• To propose an exploratory data analysis (EDA)-based approach for robust QRS detection.
• To eliminate the need for parameter selection, setting, and training in QRS detection.

Main Methods:

• ECG signal pre-processing using median and moving average filters.
• Enhancement of filtered ECG peaks by raising the signal to the third power.
• Decision-making based on estimated root mean square (rms) of the signal.
• Novel isoelectric line identification and EDA-based QRS detection strategy.

Main Results:

• Validation performed on over 1 million beats from diverse databases (MIT-BIH, FDB, ESTD, SDB, FTDB).
• Achieved high performance metrics: 99.65% sensitivity and 99.84% positive predictivity rate.
• Demonstrated a computationally efficient and parameter-free QRS detection method.

Conclusions:

• The proposed EDA-based technique offers a simple yet effective solution for QRS-complex detection.
• The method's robustness and accuracy make it suitable for real-time cardiac disorder monitoring.
• Eliminates the need for user intervention in parameter setting, enhancing clinical applicability.