Sensors & Transducers (Mar 2021)
Visual Intelligence-Based Contactless Bio Signals Measurement and Diagnosis Sensor to Overcome COVID-19
Abstract
COVID-19 is mainly transmitted either directly by droplets from infected people or indirectly through contact with vehicles in contaminated environments. Coronavirus, which is highly contagious, is highly likely to result in a secondary infection between the patient and the medical staff during the first diagnosis, and if it is left alone, the likelihood of a tertiary infection is increased. In order to prevent this, a biometric means capable of diagnosing body temperature and heart rate in a non-contact manner is needed in consideration of the characteristics of corona symptoms accompanying high fever and various complications. In this paper, the body temperature image and pulse image are acquired simultaneously using a general camera to measure the health status of the subject. Using this, it is possible to monitor in real time and to react immediately to the patient's acute disease. This study aims to develop a system capable of simultaneously estimating temperature by measuring heart rate using image sequences and measuring changes in peripheral blood flow. First, the system recognizes the user's face from a sequence of images collected through the camera, then tracks it to continuously extract the color of the same area and measures the user's heart rate using the color change of the corresponding pixel. Second, the system measures changes in peripheral blood flow due to external temperature. The body maintains normal body temperature by regulating peripheral blood flow. In this study, the user's body temperature was estimated by measuring the difference in this delay with the user's body temperature. As a result of the experiment, we confirmed that there is a close relationship between the external temperature and the stabilization time of blood flow, and based on this, a body temperature inference equation was derived. In conclusion, this study, a user's face was recognized using a webcam, and a system was developed to infer heart rate and body temperature based on the color data of the recognized user's face.
