Performance Evaluation of Ten SpO2 Measurement Equations Using Estimation Error Range Metric and Different Signal Lengths

Although many pulse oximeters are available in market with different blood oxygen saturation (SpO2) equations presented in the past studies, we investigate the performance of ten existing SpO2 equations by validating the robustness of calibrated coefficients and ratio of ratios (R) parameter. For this performance study, we used the pulse transit time (PTT-PPG) database having different dual-wavelength PPG signals with their ground truth SpO2 values. From the evaluation results in terms of mean absolute error (MAE), standard deviation (SD), root mean square error (RMSE) and Bland-Altman ratio (BAR), identifying the best SpO2 equation(s) is challenging because the above-mentioned error metrics are biased with averaging of error values with large number. Therefore, this study presents a new estimation error range (EER) metric to highlight the large error values of the SpO2 estimation algorithms. On the EER metric, the best SpO2 equation is presented based on the estimation results obtained for two different durations of signal.