Dynamical-Statistical Approach to Evaluation of Cardiac Signals

The present study has proposed a checking and filtration algorithm for processing cardiac signals – ECG and blood pressure (BP) with various frequencies. The algorithm is based on analysis of correlation matrices (principal component method) of the second derivative of signals’ rows. The method for cardiac signals approximation by smoothing cubic splines has been developed. Also the study includes the analysis of the dependence between the number of spline nodes and the norm of deviation of the spline approximation from the source values of the signal. As the result of the study a program for detecting local maxima and minima of BP signal and P-, QRS-, T-complexes of ECG signal has been developed. Previously the signal noises have suppressed by applying of the local Fourier filter and the Shuman filter. A method for discovering abnormal signal areas based on correlations between signal parameters has been proposed. The study has also proposed new characteristics of a set of measurement series for a given patient, which could be applied to identification and classification of possible functional changes or pathologies in the functioning of the cardiovascular system.