Efficient Fiducial Point Detection of ECG QRS Complex Based on Polygonal Approximation

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Accurate electrocardiogram (ECG) analysis is crucial for diagnosing arrhythmias.
• Fiducial point detection (onset, offset, peak) is essential but challenging due to ambiguous feature values.
• Existing methods struggle with the inherent ambiguity in detecting fiducial points.

Purpose of the Study:

• To enhance the accuracy of fiducial point detection in ECG signals.
• To develop a reliable method for identifying QRS complex onset and offset.
• To improve the diagnostic accuracy of arrhythmias through precise ECG analysis.

Main Methods:

• Utilized a curvature-based vertex selection technique with polygonal approximation to represent ECG signals.
• Minimized candidate samples for fiducial point detection by emphasizing key feature values.
• Generated an auxiliary signal based on accumulated amplitude change rate between vertices to capture morphological changes.

Main Results:

• Achieved stable mean and standard deviation errors for onset (-4.02 ± 7.99 ms) and offset (-5.45 ± 8.04 ms) detection.
• Demonstrated the method's effectiveness through QRS complex clustering and experiments on MIT-BIH Arrhythmia Database.
• Confirmed reliable fiducial point detection across various QRS complex morphologies.

Conclusions:

• The proposed polygonal approximation method offers a practical and accurate approach for ECG fiducial point detection.
• The technique effectively addresses the ambiguity in detecting onset and offset points, improving diagnostic reliability.
• This method shows significant potential for enhancing automated arrhythmia diagnosis systems.