Detection of fiducial points in ECG waves using iteration based adaptive thresholds

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Accurate detection of fiducial points in electrocardiogram (ECG) signals is crucial for diagnosing cardiac conditions.
• Existing algorithms may struggle with computational complexity or adaptability to varying ECG morphologies.

Purpose of the Study:

• To develop and validate an efficient algorithm for detecting fiducial points in ECG waves.
• To utilize adaptive thresholds and reference salient points for robust fiducial point identification.
• To assess the algorithm's feasibility for mobile device implementation.

Main Methods:

• An iterative algorithm employing adaptive thresholds to detect reference salient points (RSPs) within the R-peak interval.
• Fiducial point detection based on the number of identified RSPs, accommodating normal and abnormal ECG variations.
• Validation using twelve records from the MIT-BIH Arrhythmia Database with manual fiducial point annotations.

Main Results:

• The algorithm achieved mean absolute distance errors of 12.2 ms (P-wave start), 7.9 ms (Q-wave start), 9.3 ms (S-wave end), and 13.9 ms (T-wave end).
• The computational complexity was found to be very low.
• The algorithm demonstrated high accuracy comparable to manual annotations.

Conclusions:

• The proposed algorithm offers an accurate and computationally efficient method for ECG fiducial point detection.
• Its low complexity makes it suitable for real-time analysis on resource-constrained mobile devices.
• This advancement supports the development of portable ECG monitoring solutions.