Specifying respiratory noise in the ICG signal to measure hemodynamic parameters

In ICG measurement, the respiratory component is considered as a noise source targeted to be removed. Specifying the characteristics of the respiratory noise in measuring ICG signal is inevitable in order to design an effective noise filter. This study aims to introduce a method to directly extract the respiratory noise from the obtained ICG signal to truly specify the features of the noise in terms of power and frequency spectrum. The algorithm is based on the data of stop breathing and breathing states in combination with ensemble averaging technique. The algorithm uses the simultaneously recorded electrocardiography (ECG) signal as the reference to segment the data. The proposed algorithm shows great performance on three regular states of respiration with the large variation of the magnitude and frequency including normal breathing, rapid breathing, and forced breathing.