Hankel Matrix Based Algorithm for Cardiac Pulse Wave Base and Peak Lines Correction

Abstract

In the paper we proposed a new method for the cardiac pulse wave base lines and peak lines estimation and correction. The proposed method is mainly directed, but not limited, to the Electrical Bio-Impedance (EBI) and Central Aortic Pressure (CAP) signals. However, the method can be extended to other signal kinds and application fields. Definitely, the proposed method can be applied to the PPG signals and blood pressure waveforms measured from different body locations, not only central aortic pressure. The base line correction approach, instead of filtering, is selected due to the physiological peculiarities of the cardiac cycle. The minimum value of a cardiac signal, which is the diastolic blood pressure (minimum pressure in the cardiac cycle), varies much less than the systolic peak value. Thus, in our research work we use the base line correction (subtraction) instead of mean value subtraction (filtering) to get cardiac signal's component. The proposed method is based on combination of the mathematical morphology and on the Hankel matrix. The method does not need separate estimates of peaks and valleys of the waveforms. Moreover, for correctly estimated signal frequency, the proposed method estimates the base line and the peak line as a piecewise lines between signal's minima or maxima. The result is a corrected cardiac signal that does not need additional processing, based on piecewise estimates of the base and peak lines.