Analysis of Pulse Oximetry Signals through Statistical Signal Processing Techniques

Abstract

Signal processing techniques for Pulse Oximetry can typically be applied to estimate the oxygen saturation in the time and frequency domain. This paper describes whether or not Pulse Oximetry signals are consistent to calculate oxygen saturation for different level of measurements through the three patients’ data. This is accomplished using the statistical signal processing techniques; Autocorrelation Function (ACF), Crosscorrelation Function (CCF), Power Spectral Density (PSD), and Coherency. We can confirm the consistent patterns of Pulse Oximetry signals from the limited experiment and see the fact that Pulse Oximetry may not be calibrated correctly or even fail to read the oxygen saturation.