An Application of Dual-Q Tunable Q-factor Wavelet Transform for QRS Detection in ECG Signal

Abstract

Accurate localization of QRS complexes in the electrocardiographic signal is essential in clinical practice for prevention of heart disease. In this paper, we propose a new method for QRS complex detection based on the decomposition of electrocardiographic signal with dual-Q tunable Q-factor wavelet transform (Dual-Q TQWT). The proposed method starts with a preprocessing stage which consists of baseline wandering removing, and then the obtained signal is decomposed into low and high resonance components with Dual-Q TQWT. In the next stage the decomposed signal is created from the selected low resonance components. On this basis, after the non-linear mapping, the detection function waveform is derived. Applying the three-stages amplitude threshold method allows us peaks localization. These peaks correspond to locations of QRS complexes. Our approach has been evaluated over the MIT- BIH Arrhythmia Database and MIT-BIH Noise Stress Test Database. The proposed method yielded sensitivity and positive predictivity of 99.88 % and 99.86 % respectively for arrhythmia database. The results obtained are better or comparable to the state-of-the-arts methods.