Early Detection of COVID-19 through AI
Temperature sensing to help Canadians in returning to work
Project Budget* - $5.2M
Partner Co-investment - $0.7M
Supercluster Co-investment - $4.5M
As Canada and the world move to reduce the spread of COVID-19, temperature screening is increasingly used to detect potentially sick individuals. However, hand-held temperature checkpoints can create bottlenecks and hurdles.
Being able to scan large numbers of people entering hospitals, stores, restaurants, airports and train stations will be critical to protecting the health of Canadians and bending the curve of subsequent waves.
That’s the goal of Early Detection of COVID-19 through Artificial Intelligence being led by Patriot One Technologies in partnership with Cisco Co-Innovation Centre, the University of British Columbia-Okanagan, Los Angeles Football Club and the Cincinnati Reds.
The first phase involves creating a software module that works with thermal cameras connected to an existing monitoring system so individuals with elevated temperatures can be highlighted in real time. People with a fever would then be directed to secondary screening process. This can help businesses, healthcare, retail, sports, entertainment and hospitality organizations as they plan to re-open or expand operations safely.
Machine learning is at the heart of this new system. When it detects a body temperature anomaly, an alert is generated and instantly transmitted to onsite security for further action.
The next step for the project is to monitor compliance with COVID-19 safety measures such as wearing non-medical masks and physical distancing.
Real-world testing of the temperature and face mask detection modules are taking place in locations across the U.S. and Canada, including university campuses and sports stadiums. The tests will ensure the technology works in varied situations.
This multi-sensor approach also has the potential, if networked, to empower multi-location monitoring that can predict and even mitigate the impact of future outbreaks.