Model-based parameter estimation applied on electrocardiogram signal

An Electrocardiogram (ECG) feature extraction system was developed based on the calculation of the poles employing Pade’s approximation techniques. Pade’s approximation was applied on five different classes of ECG signals’ arrhythmia. Each signal was represented as a rational function of two polynomials of unknown coefficients. Poles were calculated for this rational function for each ECG signals’ arrhythmia and were evaluated for a large number of signal windows for each arrhythmia. The ECG signals of lead II (ML II) were taken from MIT-BIH database for five different types. These were the ventricular couplet, ventricular tachycardia, ventricular bigeminy, and ventricular fibrillation and the normal. ECG signal was divided into multiple windows, where the poles were calculated for each window, and was compared with the poles computed from the different arrhythmias. This novel method can be extended to any number of arrhythmias. Different classification techniques were tried using neural networks, K nearest neighbor, linear discriminate analysis and multi-class support vector machine.   Key words: Arrhythmias analysis, electrocardiogram, feature extraction, statistical classifiers.