A wavelet-based ECG delineation algorithm for 32-bit integer online processing.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Wavelet Transform (WT) has shown promise for electrocardiogram (ECG) analysis since 1995.
• Previous WT implementations for ECG delineation used computationally demanding non-linear operators.
• Efficient online processing of ambulatory ECG signals requires computationally efficient methods.

Purpose of the Study:

• To develop an advanced, computationally efficient algorithm for online QRS detection and P-QRS-T wave delineation.
• To utilize a 32-bit integer, linear algebra approach for ECG signal processing.
• To overcome the computational demands of previous non-linear WT methods.

Main Methods:

• A novel 32-bit integer, linear algebra approach based on Wavelet Transform (WT).
• Application to single-lead ECG signal processing for QRS detection and P-QRS-T wave delineation.
• Validation using established ECG databases (MIT-BIH Arrhythmia, European ST-T, QT Database).

Main Results:

• High performance in QRS detection: Sensitivity (Se) of 99.77% and positive predictive value (P+) of 99.86% on the MIT-BIH database.
• Excellent performance on the European ST-T Database: Se = 99.81%, P+ = 99.56%.
• Accurate ECG delineation with mean error below 1.5 samples for all fiducial points on the QT Database.

Conclusions:

• The proposed algorithm provides reliable QRS detection and accurate ECG delineation.
• The method's simple structure, based on integer linear algebra, enhances computational efficiency.
• This approach is suitable for efficient online processing of ambulatory ECG signals.