Remote Vital Sign Monitoring in Accidents

This paper proposes to evaluate the feasibility of obtaining the heart rate (HR) through remote Photoplethysmography (rPPG) while the subject is lying down in an emergency situation. The proposed methodology performs image and signal processing on frames containing part of a person’s skin to infer the blood volume variation signal, and then, estimate the HR. As the study evaluates the accuracy of the methodology in lying down subjects, three poses were chosen for the tests: Ventral Decubitus, Dorsal Decubitus and Lateral Decubitus. The results demonstrate the possibility to obtain reliable heart rate values in these conditions with average Root Mean Square Error (RMSE) of 3.60 bpm. Finally, it is also demonstrated that the method can be embedded in an ordinary smartphone to facilitate its use.