Adaptive R-peak Detection Using Empirical Mode Decomposition

Abstract

Accurate QRS detection plays a pivotal role in the diagnosis of heart diseases and the estimation of heart rate variability and respiration rate. The investigation of R-peak detection is a continuing concern in computer-based ECG analysis because current methods are still inaccurate and miss heart beats. This paper presents a different algorithm to the state-of-the-art Empirical Mode Decomposition based algorithms for R-peak detection. Although our algorithm is based on Empirical Mode Decomposition, it uses an adaptive threshold over a sliding window combined with a gradient-based and refractory period checks to differentiate large Q peaks and reject false R peaks. The performance of the algorithm was tested on multiple databases including the MIT-BIH Arrhythmia database, Preterm Infant Cardio-Respiratory Signals database and the Capnobase dataset, achieving a detection rate over 99%. Our modified approach outperforms other published results using Hilbert or derivative-based methods on common databases.