|
Since the outbreak of SARS (2003) / new influenza (2009) / MARS (2013) and so on, thermography-based Fever Screening has been adopted in quarantine of airports and ports as a measure to prevent from the spread of infection. However, the fever judgement has been achieved with the experienced quarantine officer. Due to recent spread of COVID 19, installation of thermography is expanding not only to quarantine at airports and ports, but also to hospitals, schools, retail stores, various facilities, and sports/event venues. Automated fever screening system with high accuracy are needed. Generally, a thermometer that measures axillary or sublingual temperature is used for body temperature measurement accurately, but this method takes time and contacts the subject to measure. So it is not suitable for the screening. On the other hand, thermography, which can measure the temperature of several persons in real time in a non-contact manner, satisfies screening requirements, but the body surface temperature of the exposed part such as the face that can be measured, is strongly affected by the environmental temperature and fluctuates. Therefore, it has a problem that it is difficult to determine the presence or absence of fever by setting a constant threshold value only. In order to realize the automation of screening with high accuracy and efficiency and make it easy for general users to understand, we will increase the measurement accuracy of the body surface temperature of thermography, identify the face and each landmark position from the image, and measure the temperature (body surface) of that part. The effects of the environment and individual difference are corrected from the temperature of multiple specific parts, and the high-precise body temperature estimation is carried out, and the algorithm is further improved.
|
