Covid-19 can be detected in minutes thanks to artificial intelligence (AI) technology developed at the University of the West of Scotland.
The program is an improvement over PCR tests that typically take about two hours.
It is hoped that the technology can eventually be used to ease the burden on emergency departments, particularly in countries where PCR testing is not readily available.
The technique uses X-ray technology, which compares scans to a database of approximately 3,000 images of patients with Covid-19, healthy individuals and those with viral pneumonia.
MORE FROM MEDICAL AND HEALTH CARE
It then uses a deep convolutive neural network, an algorithm typically used to analyze visual images, to make a diagnosis. During an extensive testing phase, the technique proved to be more than 98 percent accurate.
Professor Naeem Ramzan, director of the Affective and Human Computing for SMART Environments Research Center at UWS, led the team behind the project.
He said: “There has long been a need for a fast and reliable tool that can detect Covid-19, and this has become even more true with the emergence of the Omicron variant.
“Several countries are unable to conduct large numbers of Covid tests due to limited diagnostic tools, but this technique uses easily accessible technology to quickly detect the virus.
“Covid-19 symptoms are not visible on X-rays during the early stages of infection, so it is important to note that the technology cannot completely replace PCR testing.
“However, it can still play an important role in containing the spread of viruses, especially when PCR tests are not readily available.
“It may prove crucial and potentially life-saving in diagnosing severe cases of the virus, to help determine what treatment may be needed.”
The team — including researchers from Durham University and Northern Border University, Saudi Arabia — plans to expand the study with a larger database of X-ray images obtained by different models of X-ray machines, to assess the suitability of the approach in a clinical setting.
Published in Sensorsthe findings of the team can be found here.
