QRS Complex Detection using Empirical Mode Decomposition based Windowing Technique

In this work an Empirical Mode Decomposition based QRS complex detection algorithm is proposed. Other decomposition techniques use some predetermined basis function for transformation and hence may not be applicable for all kind of signals. Being a fully data driven adaptive technique, the present method depends on selection of proper and optimum set of IMFs to generate an intermediate signal. Some simple mathematical operations are performed on that signal to highlight the R peak. Then the Q and S points are detected by windowing technique. Only second and third IMFs are required for QRS complex detection. The proposed method is tested with PTB diagnostic database and MIT-BIH Arrhythmia database. The R peak detection success rate is 98.67%. Sensitivity and specificity of QRS complex detection is 98.88% and 99.04% respectively.