Estimating SpO2 with Deep Oxygen Desaturations from Facial Video Under Various Lighting Conditions: A Feasibility Study

This paper presents a feasibility study to collect data, process signals, and validate accuracy of peripheral oxygen saturation (SpO 2 ) estimation from facial video in various lighting conditions. We collected facial videos using RGB camera, without auto-tuning, from subjects when they were breathing through a mouth tube with their nose clipped. The videos were record under four lighting conditions: warm color temperature and normal brightness, neutral color temperature and normal brightness, cool color temperature and normal brightness, neutral color temperature and dim brightness. The air inhaled by the subjects was manually controlled to gradually induce hypoxemia and lower subjects' SpO 2 to as low as 81%. We first extracted the remote photoplethysmogram (rPPG) signals from the videos. We applied the principle of pulse oximetry and extracted the ratio of ratios (RoR) for two color combinations: Red/Blue and Red/Green. Next, we assessed SpO 2 estimation accuracy against the ground truth, a Transfer Standard Pulse Oximeter. We have achieved an RMSE of 1.93% and a PCC of 0.97 under the warm color temperature and normal brightness lighting condition using leave-one-subject-out cross validation between two subjects. The results have demonstrated the feasibility to estimate SpO 2 remotely and accurately using consumer level RGB camera with suitable camera configuration and lighting condition.Clinical Relevance— This work demonstrates that SpO 2 can be estimated accurately using an RGB camera without auto-tuning and under warm color temperature, enabling continuous SpO 2 monitoring applications that require noncontact sensing.