The work of a team at the Quebec Institute of Scientific Research (INRS) could one day lead to more effective and accurate fever detection in public places. Traditional thermal cameras measure temperature on the forehead or cheeks, which means that their measurements can be disrupted by make-up, wearing a mask, perspiration or even air circulation. Researchers led by professor Jinyang Liang have therefore developed a thermal camera called SPIRIT that focuses instead on the inner canthus—the corner of the eye where the eyelids meet, on either side of the nose. “We measure temperature in a completely unconventional way,” said Liang, who discussed his work with The Canadian Press. “We measure it from a single pixel. It’s essentially a computational imaging system that allows us to overcome existing limitations, such as those of thermal cameras.” Whether using a conventional thermometer or an infrared device to measure temperature, Liang continued, “we are measuring the temperature of an area covered by skin.” “So if you’re sweating, or if you’re wearing make-up, the measurement may not be very accurate,” he added. “Scientific literature has shown us that one of the best places for an accurate temperature reading is the inner canthus. It has been demonstrated that the temperature measured in this area corresponds more closely to core body temperature, which is essential for reflecting the physiological state of a human being.” While the inner canthus has the advantage of not being covered by skin, these two areas of the eye are tiny and far apart. The technical details are incomprehensible to the average person, but SPIRIT (for “single pixel infrared imaging thermometry”) apparently solves this problem by using a single infrared pixel, combined with a light encoding system and computer reconstruction. This approach focuses 100 per cent of detection on the most reliable areas, without the need for expensive or complex systems, according to a press release. Traditional thermal cameras, said Liang, “are like shooting a mosquito with a cannon: the vast majority of pixels are not even used, they are wasted, and the ones you do use may not give you the accuracy you need.” In contrast, Liang and his colleagues explained in the journal Nature Communications in October, “SPIRIT enabled passive mapping of human inner canthus temperature in 15 seconds, with an overall image size of 11 to 13 pixels and a temperature resolution of 0.3 °C.” SPIRIT, they add, “devotes its entire field of view to the inner canthus, which improves measurement efficiency and reduces inaccuracies in temperature readings due to background mixing.” The new camera is also able to “detect temperature differences of less than one degree related to daily physical activities, gender and wearing glasses,” which “allows for more accurate assessment criteria for public health (e.g., for COVID-19 fever screening).” The researchers also reveal that their camera unexpectedly detected the onset and development of a fever in a volunteer who initially presented with no fever. During their work, they also found that the temperature of the two canthi differs slightly; that there are differences between men and women; and that the temperature fluctuates constantly throughout the day. A few obstacles still need to be overcome before the SPIRIT camera can be deployed on a large scale. At present, for example, the subject must remain still for 15 seconds to allow the camera to take a reading, which would be far from practical in a setting such as an airport. But once perfected, the technology could be integrated into existing measures, “since when you go through customs, you have to have your photo taken anyway,” said Liang. Its deployment could also be considered one day in schools, medical clinics, railway stations or workplaces. The researchers hope that the accuracy of their camera will reduce the number of false alarms (false positive results), thereby reducing the stress and costs associated with unnecessary testing. “During the COVID pandemic, there were so many unnecessary cases, such as false positives, which actually caused significant stress for people,” Liang said. “There is a burden from a psychological, physical and even financial point of view. So we hope that (...) the next time a pandemic strikes, decision-makers will be more aware of this subtle difference, which can actually have a significant impact on individuals.” The researchers are considering applications other than temperature control, such as the detection of abnormal thermal signatures associated with certain diseases (e.g., inflammation, circulatory disorders and breast cancer) or the detection of objects in conditions of reduced visibility (e.g., fog, smoke and occlusion), according to Nature Communications. Other potential applications include infrared astronomy, atmospheric observations and the characterization of advanced materials. This report by The Canadian Press was first published in French on Nov. 27, 2025.
