Data-Driven Calibration Estimation for Robust Remote Pulse-Oximetry

Abstract

Pulse-oximetry has become a core monitoring modality in most fields of medicine. Typical dual-wavelength pulse-oximeters estimate blood oxygen saturation (SpO 2 ) levels from a relationship between the amplitudes of red and infrared photoplethysmographic (PPG) waveforms. When captured with a camera, the PPG waveforms are much weaker and consequently the measurement is more sensitive to distortions and noises. Therefore, an indirect method has recently been proposed where, instead of extracting the relative amplitudes from the individual waveforms, the waveforms are linearly combined to construct a collection of pulse signals with different pulse signatures, each corresponding to a specific oxygen saturation level. This method has been shown to outperform the conventional ratio-of-ratios based methods, especially when adding a third wavelength. Adding wavelengths, however, complicates the calibration. Inaccuracies in the calibration model threaten the performance of the method. Opto-physiological models have been shown earlier to provide useful calibration parameter estimates. In this paper, we show that the accuracy can be improved using a data-driven approach. We performed 5-fold cross validation on recordings with variations in oxygen saturation and optimized for pulse quality. All evaluated wavelength combinations, also without visible red, meet the required ISO standard accuracy with the calibration from the proposed method. This scalable approach is not only helpful to fine-tune the calibration model, but even allows computation of the calibration model parameters from scratch without prior knowledge of the data acquisition details, i.e., the properties of camera and illumination.

Editorial for Special Issue: Contactless Vital Signs Monitoring

“…Nevertheless, it has been shown to be possible, even during some patient motion [6]. In their current paper [7], they demonstrate an elegant data-driven way to perform calibration in a less cumbersome way than earlier calibration efforts [8]. In a pioneering contribution by Lorato et al [9] a non-camera sensor, a thermopile array, was demonstrated to be capable of monitoring respiratory flow.…”

Human Vital Signs Detection Methods and Potential Using Radars: A Review

“…The green light is widely used for HR acquisition. However, many photoplethysmography setups use the red and infra-red optical regions for blood oxygen monitoring [38][39][40]. A portion of the emitted light is absorbed by human tissue while the other portion is reflected.…”

A review of wearable and unobtrusive sensing technologies for chronic disease management

“…The conventional signal combination method is the “ratio-of-ratios (RRs) method based on LBL [ 132 ]. In previous studies, one research direction has been to replace the RRs method with other methods, such as combining the obtained PPG signals to construct a pulse signal set with the best signal quality [ 133 ], [ 134 ]. Further studies have also attempted to replace the combination of red and near-infrared light with a combination of visible lights, such as red and blue light [ 135 ], or red and green light [ 122 ], with the intention of remotely monitoring SpO 2 using consumer cameras, such as those on smartphones.…”