Heart Rate Monitoring With Smartphone Built-In Frontal Digital Camera

Non-contact technologies are gaining much interest as promising systems for remote monitoring of physiological parameters (e.g., heart rate -HR-, respiratory rate, blood pressure, blood oxygen saturation) without interfering with the subject's comfort. Among the existing technologies, digital cameras integrated into smartphones or laptops are widely used due to their ease of use, availability, and portability. In this study, we investigated the influence of distance on the HR estimation from a video recorded with a smartphone's frontal camera. HR values were provided from a spectral analysis every 1 s. We evaluated the performances of six different algorithms to identify the best one for HR estimation simulating an occupational scenario at two distances (i.e., 0.5 m and 1 m). Data were recorded from 8 healthy volunteers of both sexes; a wearable device was used to record medical-grade ECG to estimate reference HR values. Results show that the higher the camera distance, the higher the mean absolute error (MAE): the average MAE was 1.49 bpm at 0.5 m and 2.59 bpm at 1.0 m. Moreover, Green Channel, Chrominance-based signal processing method (CHROM), and Plane Orthogonal to Skin (POS) have been identified as the best algorithms to estimate HR since they showed a better agreement with the reference system than the other algorithms.