“For coronavirus test, cough into the microphone”



An artificial intelligence algorithm developed at MIT accurately identifies patients with Covid-19 by analyzing the sound of their cough. Research team at the Massachusetts Institute of Technology reports that the system identified 98.5% of participants who were tested in a SARS-CoV-2 coronavirus test, a rate that skyrocketed to 100% for confirmed asymptomatic cases. if we remain asymptomatic “comments in a press release Brian Subirana, a member of the research team. In fact, the researchers point out that the changes in cough that the algorithm recognizes are not perceived by human hearing. in Medicine and Biology, comes as a follow-up to previous research on the diagnosis of Alt The MIT algorithm is another example of machine learning, an approach that teaches machines to perform a task by supplying them with thousands of examples. than the radiologists themselves. In this case, the researchers managed to collect 70,000 audio files with cough sounds, of which 2,500 came from confirmed symptomatic and asymptomatic cases. “This is the largest cough database we know of,” boasts Dr. Subirana. According to the BBC, the Cambridge database contains just 459 audio files, and the success rate in diagnosing Covid-19 is limited to 80%. Researchers are now working to develop an app that will allow users to cough into its microphone According to Dr. Subirana, such an application “would drastically reduce the spread of the pandemic if everyone used it before entering the classroom, factory or restaurant.” Google News and be the first to know all the newsSee all the latest News from Greece and the World, at



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *