Accurate detection and classification of ventricular abnormalitiesby using of morphological features

Electrocardiogram (ECG) is a bioelectrical signal which presents theelectrical activity of different parts of heart in time domain. Theaccurate study of this signal leads to precise detection of ventriculararrhythmias that are serious problem in the world and might lead toSudden Cardiac Death (SCD). This research attempted to achieve theaccurate and fast detection and classification of ventricularabnormalities from morphological features by using cross correlationbetween normal sinus beat and abnormal beats. Some of these featuresfor normal sinus rhythm and acute ventricular abnormalities are QSinterval and amplitudes of R. Databases of healthy subjects, arrhythmiapatients and T wave alternans were given from Physionet’s normalsinus rhythm, sudden cardiac death database and TWA-MIT,subsequently.T wave alternans is small changes in amplitude, duration or phase of Twaves in consecutive beats. Cross correlation between RR intervals andTT intervals in zero lag was used because of separation betweennormal sinus rhythm and T wave alternans. Specific morphologicalfeatures were attained for T wave alternans and normal sinus rhythmsignals. The classification accuracies of normal sinus rhythm withventricular tachycardia, ventricular flutter, ventricular fibrillation,premature ventricular contraction and escape beat rhythm were obtained100%, 92.8%, 86.8%, 90.81%, 96.28% and 91.8%, subsequently.Classification accuracy of T wave alternans was achieved 95.03%.

Keywords:

electrocardiogram , morphological features , sudden cardiacdeath , ventricular arrhythmias