Analysis of instantaneous cardiac EBI signal variability over the heart cycle(s): Non-linear time-scale approach

Abstract

The paper presents a method similar to the dynamic time warping (DTW) that normalizes cardiac Electrical Bioimpedance (EBI) periods’ waveforms in time domain by scaling period waveforms along the time axis. In opposite to DTW, where the matrix of local distance measures is analyzed to find the best warping path, the proposed method uses adaptive Bézier curve fitting approach to find “warping path”, minimizing the sum of squared error measures between the EBI periods waveforms. For the normalization, the vector of time instances of the cardiac period is approximated by the Bézier curve with 6 points (4 control points and 2 end points). Results are presented in time domain for 2 persons selected from the available patients population of 53 voluntary persons.