Signal processing using Fourier & wavelet transform for pulse oximetry

Current pulse oximeters use a weighted moving average technique to compute oxygen saturation (SpO/sub 2/) values. This method has many limitations including susceptibility to motion artifact, background light, and low perfusion errors. The goal for developing an alternate method for pulse oximetry was to overcome these limitations. The hypothesis was that frequency domain analysis could more easily extract the cardiac rate and amplitude of interest from time domain signal. The focus was on the digital signal processing algorithms that had potential to improve pulse oximetry readings, and then test those algorithms. This was accomplished using the fast Fourier transform (FFT) to analysis in spectral domain, and the discrete wavelet transform (DWT) to estimate oxygen saturation. The result indicate that the FFT and DWT computation of oxygen saturation were accurate and erroneous without weighted moving average (WMA) algorithms currently being used.