Realtime Detection of ECG Signal R-peaks Using a Lightweight RREADER Algorithm

Abstract

The accurate identification of the R-peak of an electrocardiogram (ECG) cycle is the first step towards automated classification of the cycle. In this paper we present a lightweight algorithm, known as R-READER: a Rapid-Ramp Effective Algorithm for Detection of ECG R-peaks. It is based on an intuitive and effective identification of inflexion points in the ECG cycle through the calculation of adjacent slopes between the signals. It is compared with the Pan-Tompkins (P-T) algorithm a popular algorithm used for ECG QRS detection. We have used the best (#100) and worst (#108) cases in the MIT-BIH Arrhythmia Database for our analysis and had average accuracy, positive predictivity and sensitivity values of 97.79%, 98.23% and 99.54% for worst-case scenarios respectively when compared to values of 87.47%, 88.57% and 98.59% for the P-T algorithm, all only at a fraction of the latency and resource demands.