Automated QRS Detection using Empirical Mode Decomposition and K-Means

This paper proposes an algorithm using Empirical Mode Decomposition (EMD) and k-means for the detection of QRS complexes present in the ECG signal. EMD is an innovative method for decomposing any time varying, nonlinear and non-stationerysignal into a set of intrinsic mode functions (IMF). This automated algorithm is applied to the filtered ECG signal for its decomposition into its intrinsic components and further its classification is done using k-means classifier. Dataset-3 of the CSE multi-lead measurement library is used for validating the performance of the algorithm. Detection rate of the proposed algorithm came out to be 99.42% with sensitivity (Se) and prediction (+P) rates being 99.39% and 99.93% respectively. The performance of this algorithm is quite satisfactory amongst many algorithms used for the automated detection of QRS complexes