부산대학교 도서관

The detection of cardiac diseases is very important to prevent early cardiac death. The electrocardiogram (ECG) signal is an informative as well as non-invasive clinical tool to diagnose such diseases of the human heart. In this paper, a MATLAB based graphical user interface (GUI) has been developed, and scrupulous attention has been paid for cardiac signal processing to distinguish the subject with malignant ventricular ectopy (MVE) from one having normal sinus rhythm (NSR). Here, long-duration ECG segments of 60 seconds have been considered from the MIT-BIH database, and different analytical approaches have been conducted throughout the study. In heart rate variability (HRV) analysis, significant variations of heart rate have been observed for the subjects with MVE compared to NSR. Afterward, from linear analysis, the extracted time-domain and frequency-domain features of each subject provide divergent numerical values to differentiate these two classes effectively. Also, the numerical values of these features have been displayed in GUI which provides an overall picture of a subject’s heart condition in a single window. Moreover, time-frequency analysis of ECG signals has been carried out using continuous wavelet transform (CWT) that also categorizes particular heart diseases significantly. In addition, the incorporation of a convolutional neural network (CNN) in the GUI is a novel approach that facilitates the arrhythmia detection and classification process. Here, the CWT scalogram image of ECG signal of a particular subject has been used as input in CNN which identifies the subject suffering from MVE or not with 100% accuracy.