Low-Cost Multi-Person Continuous Fever Screening

Xiaofan (Fred) Jiang, Assistant Professor of Electrical Engineering and Computer Engineering, Columbia University Department of Engineering 

This project proposes a thermal vision-based low-cost system for continuous detection of fever in large populations to combat the spread of infectious diseases such as COVID-19. One of the key challenges in containing infectious diseases such as COVID-19 from spreading is the lack of a safe and cost-effective way to perform early detection of potential carriers at scale. It is particularly damaging when the virus has a long incubation period, rendering 1-shot checks, such as those performed at airports, largely ineffective. What is desperately needed is an inexpensive way to continuously screen large populations, such as in hospitals, workspaces, classrooms, buses, cafes, and homes. This project aims to use low-cost (~$300) thermal cameras to perform continuous skin temperature measurements, enabling anyone to perform fever screening remotely and safely. Our system is composed of six main subcomponents. The different subcomponents connect to extract features from the RGB and thermal images, and output fever predictions for all occupants in the field of view. Thermal and RGB image features are fused into 3D point-cloud models of heads to improve accuracy.