Infrared fever body identification using shape and temperature filters

Infrared cameras working in the range of long wave infrared region from 8 to 14 μm are used in fever screening. An algorithm to automatically detect human subjects with fever in infrared images is presented in this paper. The algorithm was tested on a training data population to obtain a decision threshold in the framework of Neyman-Pearson hypothesis testing. Results show a high detection frequency. The algorithm is illustrated by testing it with simulated images and with images acquired by an infrared camera system. Several relevant scenarios involving different radiating objects in the field of the camera were reproduced and tested.