Evaluation of two QRS detection algorithm on ECG stress test database

Abstract

In this paper, we evaluated two well-known QRS algorithms: Pan & Tompkins (PT) and based wavelet transform (WT) on an ECG stress test database. In the absence of an annotated ECG stress test database, the first stage of this work consisted of the database annotation, using RR-time series obtained from an eight leads stress database (DICARDIA). First, the system proposes to users a lead (reference channel) according to its statistical measures. Then the user realizes a visual inspection aimed at validating or denying the channel proposed by the system. As the series contains few artifacts, the annotation is performed using interval of annotations. Preliminary results realized over 31928 beats provide a sensibility of 99.81% and 98.28% respectively for PT and WT. The procedure developed in this work can be seen as a valuable starting point in semiautomatic annotation of large electrocardiographic databases, as well to evaluate and to improve stress ECG delineations.