Abnormal ECG Signals Analysis Using Non-Parametric Time–Frequency Techniques

Abstract

Due to the high mortality rate of heart diseases, early detection and precise discrimination of ECG arrhythmia are essential for the treatment of patients. Biomedical signals, especially the electrocardiogram (ECG) signal, contain information of the anatomical and physiological state of the human body. The non-stationary multicomponent nature of ECG signals makes the use of time–frequency analysis inevitable. Time–frequency signal analysis offers simultaneous interpretation of the signal in both time and frequency, which allows local, transient or intermittent components to be elucidated. The choice and selection of the proper time–frequency technique that can reveal the exact multicomponent structure of the ECG signals, especially the QRS complex, is vital in many applications, including the diagnosis of medical abnormalities. In this work, we have applied four time–frequency techniques for analyzing abnormal ECG signals. These time–frequency techniques are the Wigner–Ville distribution, the Choi–Williams distribution, the Bessel distribution and the Born–Jordan distribution. The abnormal cardiac signals were taken from a patient with supraventricular arrhythmia and a patient with malignant ventricular arrhythmia. The results obtained showed that the Choi–Williams time–frequency technique has a superior performance, in terms of resolution and cross-term reduction, as compared to other time–frequency distributions.