Scottish software service provider Cloudsoft has helped the UK Government implement a Covid-19 test that can accurately detect the virus using only voice data.
The ‘Cough-In-A-Box’ application uses artificial intelligence (AI) to diagnose the virus through audio recordings of a user coughing directly into their smartphone’s microphone.
Analysing more than 70,000 recordings submitted by volunteers, which included around 2,500 from people confirmed to have Covid-19, researchers from MIT were able to train the AI model to correctly identify 98.5% of positive cases, including 100% of asymptomatic cases.
Since December 2020, a team of Cloudsoft engineers have been working with specialists from Fujitsu, Formwize and the Department of Health and Social Care to convert the application to the cloud.
This has allowed the UK Government to rapidly scale the application for 40,000 further trials, as well as ensure it meets the strict demands for data processing and security, due to the need of anonymising personal voice data, combining it with participant’s medical history and PCR test data from all four home nations across the UK.
In addition, by utilising a combination of Amazon Web Services (AWS) and Microsoft Azure cloud platforms, the application benefits from increased resilience and reliability, which are key requirements for the future use of the project as the UK Government expands its ongoing trials.
Ross Gray, chief executive at Cloudsoft, said: “This is transformational technology for the greater human good and we’re proud to at the forefront of technological innovation to contribute to overcoming Covid-19.
“With so many applications in the real world, this technology could be a game-changer for so many sectors – including travel and hospitality, providing the current trials continue to show the same level of effectiveness as the initial MIT study.”
David Cairns, head of innovation at Fujitsu, said: “We relied on Cloudsoft’s expertise to co-architect the application to be cloud-native, ensuring that it could meet the demands of scale, reliability and security.
“This cutting-edge approach to Covid-19 detection has huge potential and is a perfect example of innovating for the human good.”
In October 2020, researchers at the MIT revealed that they had successfully detected cases of Covid-19 in asymptomatic individuals simply by processing recordings of an individual’s cough. Their discovery opened the possibility of identifying potential coronavirus cases through an individual’s mobile phone without the need for immediate testing.
Following the successful completion of the project by Fujitsu, Formwize and Cloudsoft, the UK Joint Biosecurity Centre is now working with the Alan Turing Institute to assess the wider feasibility of the application, with the initial trial being expanded to include those taking part in the REACT1 Covid-19 prevalence survey.
Don’t miss the latest headlines with our twice-daily newsletter – sign up here for free.
