Determination of characteristic points of electrocardiograms using multi-start optimization with a wavelet transform

Abstract

The description of the main steps of the method for determination of the coordinates of the extremums of non-stationary periodic signals is given. This method is based on multi-start optimization using the wavelet transform. The main steps of the base form of multi-start optimization method with using the wavelet transform are given. The results of investigation of noise stability and error for the search of extremums of asymmetric and multi-modal test functions for such method are given. The main steps of extremum search by such method in new method for determination of the coordinates of the extremum of non-stationary periodic signals are implemented. This method is implemented for automated electrocardiograms (ECG) diagnostic systems in tele-medicine. This method allowed us to determine characteristic fragments coordinates for electrocardiogram. The procedure for esti-mation of the characteristic fragments coordinates and intervals between them is based on this multi-start optimization method with using the wavelet transform. The main steps of this procedure are described. The error in estimating the duration of the intervals between ECG characteristic fragments was estimated and the noise immunity of such estimation with increasing the noise level was evaluated. The relative error in estimation of the intervals duration between characteristic fragments was less than 4% in the case of the signal-to-noise ratio in amplitude up to 10. These results allow recommending the developed method for implementation in in-formation technologies for automated decision support systems, including telemedicine, in condition of increasing noise level in ECG signals. For further research, it is planned to develop a methodology for estimation the remaining parameters of characteristic fragments and complexes in ECG, reducing the edge effects during the estimation of the extremums coordinates.