Finding morphology points of electrocardiographic signal waves using wavelet analysis

TL;DR

This paper introduces a new wavelet-based algorithm for accurately identifying key points and wave morphology in ECG signals across multiple leads, outperforming existing methods in sensitivity and predictive value.

Contribution

The novel algorithm improves ECG wave point detection accuracy and enables wave morphology analysis, surpassing prior techniques in sensitivity and predictive performance.

Findings

Sensitivity above 97% for wave peaks

96% sensitivity for onsets and offsets

Segmentation errors below standard tolerances

Abstract

A new algorithm has been developed for delineation of significant points of various electrocardiographic signal (ECG) waves, taking into account information from all available leads and providing similar or higher accuracy in comparison with other modern technologies. The test results for the QT database show a sensitivity above 97% when detecting ECG wave peaks and 96% for their onsets and offsets, as well as better positive predictive value compared to the previously known algorithms. In contrast to the previously published algorithms, the proposed approach also allows one to determine the morphology of waves. The segmentation mean errors of all significant points are below the tolerances defined by the Committee of General Standards for Electrocardiography (CSE).