Real time implementation of arrhythmia classification algorithm using statistical methods

Cardiac ailments like Ventricular Tachycardia (VT) and Ventricular Fibrillation (VF) are life threatening conditions which require prompt medical treatment. Automatic External Defibrillator (AED) is a device that delivers shock when it encounters VF and rapid VT signals. In this paper we have attempted to develop an algorithm for AED that detects normal sinus rhythm, Ventricular fibrillation and ventricular tachycardia with minimum hands on interval. The features are extracted from the slope, time and frequency domain to classify the rhythms. The final sub algorithm calculates heart rate and distinguishes Rapid VT from slow VT. This final step is essential due the fact that when a rapid VT is encountered, shock has to be given in order to restore the normal rhythm. ECG signals from MIT and Creighton University Database have been used for the testing of algorithm. The algorithm has been successfully implemented in the DSK 6713 kit.