A Bristol-based company has devised mobile symptom-checking technology with support from Innovate UK’s COVID FastTrack funding programme.
Spookfish Innovations’ ‘fever screening kiosk’ was designed for use in public areas. The first kiosks are already in use at Bangalore international airport.
The system operates 24 hours a day and takes into account changes in ambient temperature.
Precise temperature checking
It is more precise than the widely-used ‘temperature guns’ that have been pointed at people’s foreheads during the pandemic.
Instead, Spookfish’s Harlequin thermal imaging technology uses machine learning to ‘read’ a face and find each eye’s inner tear ducts, which are close to a major blood vessel.
Subjects stand within two metres of a screen displaying an infrared image of their head. The process takes 50 milliseconds and temperature matching is achieved in less than 10 milliseconds, precise to 0.25 degrees Celsius.
Testing for the symptom, not the virus
Director Jouni Ronkainen said:
This is not a COVID-19 test. It screens for the main symptom of COVID-19, which is a fever.
Being fast and precise, it’s designed to deal with a high throughput of people. We have one at a pharmaceutical plant that consistently screens 900 employees in an hour.
The system is already installed in a number of non-public settings in the UK, most visibly at the Worcester Warriors rugby club headquarters and training centre. Another is in a high-security workforce setting.
The £50,000 grant awarded to Spookfish in Innovate UK’s COVID FastTrack funding programme allowed the company to develop a robust and rugged kiosk version of Harlequin for use in public places.
Jouni added:
The ‘ruggedised’ version can be deployed anywhere and could be a fever-screening option for when fans are finally allowed into sports stadiums.
The Innovate UK funding also helped Spookfish develop a web interface that allows multiple systems to be monitored remotely in real-time. That facility is in operation at Worcester Warriors.
Spookfish was founded in 2015. The company now employs 30 people and uses computer vision technology and machine learning techniques to solve complex problems.
Last updated: 23 March 2021