Electrocardiogram synthesis using a Gaussian combination model (GCM)

In this paper modifications to an algorithm for electrocardiogram (ECG) synthesis based on a combination of Gaussians to fit real ECG data have been proposed. A method is proposed for fitting algorithm assuming that constituent Gaussian functions in GCM model are independent. Desired period(s) of ECG were selected and the number of Gaussians in the morphologic model was determined. For ECG synthesis, a Gaussian was fitted around each of the extrema and minimized local error that is defined as local difference of real ECG and our model. The range of Gaussian fitting (place to put independent Gaussian) was determined using two methods: zero crossing method and minimum bank method. Results were presented based on the efficiency of determining the Gaussian parameters in terms of time for fitting and accuracy of model.