Combined entropy based method for detection of QRS complexes in 12-lead electrocardiogram using SVM

Area of Science:

• Biomedical Signal Processing
• Cardiovascular Diagnostics
• Machine Learning in Healthcare

Background:

• Accurate QRS complex detection is crucial for analyzing electrocardiogram (ECG) signals.
• Traditional methods face challenges with noise, including power line interference and baseline wander.
• Existing QRS detection algorithms have varying performance levels on standard databases.

Purpose of the Study:

• To propose a new method for QRS complex detection in 12-lead ECG signals.
• To utilize signal entropy and Support Vector Machine (SVM) for enhanced detection accuracy.
• To evaluate the algorithm's performance against established QRS detectors.

Main Methods:

• Application of digital filtering to remove power line interference and baseline wander.
• Enhancement of QRS complexes using a Combined Entropy criterion.
• Classification of QRS and non-QRS regions using a Support Vector Machine (SVM).

Main Results:

• The proposed algorithm achieved a QRS detection rate of 99.93% on the CSE ECG database.
• Reported false positive and false negative rates were 0.54% and 0.06%, respectively.
• Superior performance was demonstrated compared to other QRS detectors tested on the same dataset.

Conclusions:

• The signal entropy-based method with SVM offers a highly accurate approach for QRS detection.
• The algorithm effectively handles noise commonly found in ECG recordings.
• This method represents a significant advancement in automated ECG analysis and cardiovascular diagnostics.