Contactless Core-temperature Monitoring by Infrared Thermal Sensor using Mean Absolute Error Analysis

Abstract

Aims and Objectives: Usually, the increase in temperature of an individual indicates the possibility of being infected with a disease that might be risky to other people, such as coronavirus. Traditional techniques for monitoring body core-temperature require body contact either by oral, rectum, axillary, or tympanic means, which are unfortunately considered intrusive in nature as well as causes of contagion. Therefore, sensing human core-temperature non-intrusively and remotely is the objective of this research. Background: Nowadays, increasing the level of medical sectors is a necessary target for research operations, especially the development of integrated circuits, sensors, and cameras, to make life easier. Methods: The solution is proposed as an embedded system consisting of the Arduino microcontroller, which is trained with a model of Mean Absolute Error (MAE) analysis for predicting Contactless Core-Temperature (CCT), which is the actual body temperature. Results: The Arduino microcontroller was connected to an Infrared-Thermal sensor named MLX90614 as an input signal and was connected to the LCD to display the CCT. To evaluate the proposed system, experiments were conducted on 31 subjects, and contactless temperature from the three face sub-regions was sensed, including forehead, nose, and cheek. Conclusion: Experimental results demonstrated that CCT could be measured remotely from the human face, including three face sub-regions, among which the forehead region should be preferred (a smallest error rate of 2.3%), rather than nose and cheek regions (2.6 % and 3.2% error rate, respectively) for CCT measurement due to the lowest error rates achieved.