Fig. 1: Smartphone-based monitoring of two key vital signs: heart rate (HR) and respiratory rate (RR). | Communications Medicine

Fig. 1: Smartphone-based monitoring of two key vital signs: heart rate (HR) and respiratory rate (RR).

From: Prospective validation of smartphone-based heart rate and respiratory rate measurement algorithms

Fig. 1

Setup of how measurements are taken: with the finger over the rear-facing camera for HR (upper panel) and using a video of the participant via the front-facing camera for RR (lower panel). Study design: to ensure generalization across skin tones for HR (n = 95) and generalization to participants with chronic respiratory conditions (chronic obstructive pulmonary disease and asthma) for RR (n = 50). Skin-tone subgroup 1 (n = 31) corresponds to Fitzpatrick skin types 1–3 (very light, light, and intermediate); subgroup 2 (n = 32) corresponds to types 4–5 (tan and brown); and subgroup 3 (n = 32) corresponds to type 6 (dark). RR study included healthy participants (n = 10) and participants with chronic respiratory conditions (n = 40). Metrics: the main measurements were mean absolute percent error (MAPE) for HR and mean absolute error (MAE) for RR. In the boxplots, the orange lines and box edges indicate the quartiles; whiskers indicate 1.5 times the interquartile range beyond the upper and lower quartiles; dots indicate individual data points (average percent error or absolute error). For the HR study, five outlier data points in the “overall” group extend beyond the axes (>10%) and are not shown; these outliers are distributed across the three skin tone subgroups (1, 2, and 2, respectively). All data points are shown for the RR study.

Back to article page