AI technology developed to diagnose covid 19 by listening to the sound of coughing.
In Australia, AI technology has been developed to diagnose covid19 by listening to the sound of coughing. The researchers said that their algorithms could be further improved to become part of a mobile phone app that diagnoses the disease.
"We've overcome a major hurdle in developing a reliable, easily accessible and early code evaluation tool that will help these individuals slow down the spread of the virus," he said. In which the symptoms do not appear.
He said that a mobile app would get rid of the worry of spreading the disease or conducting code tests at the community level, a tool needed to prevent the epidemic.
He added that it would be an important tool for regions facing a shortage of medical equipment, testing specialists and personal protective equipment. The study also found that the method could be used for other respiratory diseases. This is not the first time that an AI algorithm has been developed to identify cough code, but the Australian model has surpassed other models and will be more practical in different regions.
Researchers say previous efforts to develop such technologies, such as those at MIT and Cambridge University, have relied on data to train the AI model. It requires special details from experts, which makes it expensive and time consuming. He said that in comparison, training algorithms using targeted data such as cough samples from a single hospital or region is out of the laboratory. Performs limited performance.
"Not being able to go beyond a limit has so far been a challenge for the practical application of this technology. Our work has been successful in overcoming this problem and a methodology has been developed for training algorithms in which these labels Sound data was used. The research team trained the algorithm using the data platforms of 2 platforms COVID-19 Sound app and Koswara so that they can learn to train themselves. Now the research team to finalize this technology. And looking for potential partners to deliver apps.