Remote sensing of human skin temperature by AI speckle pattern analysis

Analysis of dynamic differential speckle patterns, scattered from human tissues illuminated by a laser beam, has been found by many researchers to be applicable for noncontact sensing of various biomedical parameters. The COVID-19 global pandemic brought the need for massive rapid-remote detection of a fever in closed public spaces. The existing non-contact temperature measurement methods have a significant tradeoff between the measurement distance and accuracy. This paper aims to prove the feasibility of an accurate temperature measurement system based on speckle patterns analysis, enabling the sensing of human temperature from an extended distance greater than allowed by the existing methods. In this study, we used speckle patterns analysis combined with artificial intelligence (AI) methods for human temperature extraction, starting with fever/no fever binary classification and continuing with temperature measurement at higher resolution.