Abstract: Ventricular Tachycardia is an abnormal heart rhythm that initiates in the ventricles. Non-sustained VTach lasts for few seconds and Sustained VTach lasts for few minutes or even hours. Sustained VTach is dangerous compared to Non-sustained VTach and if it is not treated, it often progresses to Ventricular Fibrillation. VTach is serious in people suffering with heart disease and is associated with more symptoms than other types of arrhythmia. Accurate prediction of Ventricular Tachycardia can be obtained by observing the changes in T wave, ST segment and QT interval which are the indicators for cardiac instability. In this paper, we present an algorithm that detects Ventricular Tachycardia by using morphological features of electrical signal of ventricles activity obtained from ECG. Classification of features is carried out by using Support Vector Machines (SVM). The proposed algorithm is tested on 22 records including Normal Sinus Rhythm and Ventricular Tachycardia which are collected from MIT-BIH Normal Sinus Rhythm database and CU Ventricular Tachyarrhythmia database respectively and satisfactory result is obtained as the 92.31% Specificity, 100% Sensitivity and 95.45% Accuracy is obtained.

Keywords: Normal Sinus Rhythm, Ventricular Tachycardia, Adaptive Threshold, Hilbert Transform, SVM.