Cover Story 16 IN FOCUS The system (left screen) has been deployed at Hong Kong border control points and government facilities. The Smart Fever Screening System can accurately detect a person’s temperature even when their face is partly hidden. Fever screening system safeguards public health s the COVID- crisis started to unfold, Prof. Richard SO, Industrial Engineering and Decision Analytics, immediately drew together a multidisciplinary team to devise technology to help prevent its spread in Hong Kong. Prof. So, whose research interests lie in bioinspired signal processing and auditory and visual perception, and his School of Engineering co-researchers, with expertise in machine learning, bioengineering, and parallel networking, came up with a solution in just six days. They devised a novel screening system to identify people with a fever, a signi cant symptom of COVID- and other infectious diseases, especially those moving within larger groups of travelers or passers-by. With the support of the Hong Kong government’s Electrical and Mechanical Services Department, in January , of the team’s Smart Fever Screening Systems were delivered for deployment at major border points, such as Hong Kong International Airport and Lo Wu, and later used in government facilities. The screening system was also introduced at the HKUST campus. While thermal imaging detectors have been widely used at border points for screening inbound travelers since the Severe Acute Respiratory Syndrome (SARS) outbreak hit Hong Kong, control point o cers have had to monitor two screens with separate thermal and color CCTV images to detect a sick person, all within the few seconds it takes for them to walk by. Using arti cial intelligence, real-time tracking, and decision analytics, the HKUST system more e ciently combines these two sets of images on to the same screen, with an unwell person bracketed in red and an alert triggered on detection. The new system is built on research for a completed big data platform for smart transportation, supported by the Innovation and Technology Fund and Thales Group in , Prof. So said. Through deep learning and the science of anthropometry, the fever screening system can track a face, even if three-quarters concealed. Temperature detection is more accurate as it relies on computer algorithms instead of human judgement. The system also has the capability to screen to people at the same time from a distance of up to meters. As the tracking focuses on the face, detection results are unlikely to be a ected by any heat-emitting objects that a person is carrying. In addition, the system can learn over time to become smarter and even more accurate. Researchers involved in creating the technology include Prof. Bertram SHI and Prof. Albert WONG, both Electronic and Computer Engineering, and Prof. CHEN Qifeng, Computer Science and Engineering. Other potential locations for using the system include hospitals, schools, elderly homes, and libraries. A
RkJQdWJsaXNoZXIy NDk5Njg=