Distinction of ventricular fibrillation and ventricular tachycardia using cross correlation

The accurate discrimination of ventricular tachycardia (VT) from ventricular fibrillation (VF) is an important issue in automated external defibrillator (AED) and other cardiac monitoring systems. The correlation function of the ECG signal is an adequate tool for detecting irregularities or noise in signals. Therefore, methods for VF/VT discrimination based on the autocorrelation of the ECG signal and on the cross correlation of the ECG signal with different templates have been proposed. Our work is focused on the use of the cross correlation of the ECG signal with a segment of the same ECG signal, instead of using a given template. The developed algorithm has been tested on a database consisting of 179 human VF and 74 human VT records. The new algorithm classifies accurately 313 VF windows (90.2% sensitivity) and 179 VT windows (96.75% sensitivity). These results improve those obtained from other techniques, considered as a reference, for the same records database.