Adaptive threshold QRS detection algorithm for ambulatory ECG

Abstract

The authors have developed an algorithm for detection of QRS complexes of ambulatory electrocardiogram (ECG) signals. The threshold, which is obtained from a distribution function of the amplitude of the filtered ECG signal, changes with time, adapting to changes in the QRS morphology, levels of noise, and artifacts. The threshold extracts the period which include the QRS complex, and then the QRS complex is detected during this period. By using 20 series of single-channel ECG data selected from the AHA arrhythmia database, the usefulness of this algorithm was investigated, counting both numbers of false positive (FP) and false negative (FN) QRS complexes. Correct detection of over 99% was achieved.