[ad_1]
Image source,fake images
After becoming infected with the new coronavirus, the sound of your cough will be different
Researchers in the United States have developed a clever algorithm that can tell whether a person has been infected with the new coronavirus whenever they hear the sound of a cough.
The researchers said that among people who tested positive for the new coronavirus, the success rate for this algorithm was 98.5%. Among asymptomatic patients, the accuracy rate is 100%.
This algorithm that uses artificial intelligence was developed by the laboratory of the Massachusetts Institute of Technology (MIT). The researchers published articles in the journal Engineering in Medicine and Biology of the Institute of Electrical and Electronic Engineering (IEEE).
The research team hopes to apply this technology to the mobile application for more people to use, but before that, it must be approved by the competent authority.
Image source,SCIENCE PHOTO LIBRARY
Artificial intelligence helps identify subtle differences in sound
Subtle differences
One of the article’s co-authors, MIT scientist Brian Subirana told the BBC: “After contracting the new coronavirus disease, even if you have no symptoms, the way you make your voice will change. “
He said this is as if some people cannot distinguish between men and women or children or adults by the sound of coughing. After analyzing a large number of samples to discover the subtle and special characteristics of the sound, artificial intelligence can tell whether the cough alone Infected with the new coronavirus disease.
He said this technique is particularly useful for detecting asymptomatic patients, because human ears cannot distinguish subtle differences in the cough of asymptomatic patients.
“With the reopening of schools and public transportation, practical uses for this technology can include schools, workplaces and public places, where it can be used for early warning and detection.”
Mobile phone program
It is understood that various teams, including the University of Cambridge, Carnegie Mellon University and Novoic, a British health and medical company, are also undertaking similar research projects.
The New Coronavirus Disease Sound Project at the University of Cambridge attempted to identify confirmed cases of the new coronavirus from the sounds of breathing and coughing. This project achieved an 80% success rate in July of this year.
The project collected 459 cough and breath sound samples from 378 public participants in May. Now the number of samples for the project has increased to 30,000 sound samples.
However, the MIT lab collected up to 70,000 sound samples. Each sample contained various cough sounds, of which 2,500 were from people who were diagnosed with the new coronavirus disease.
artificial intelligence
Artificial intelligence expert Calum Chace said this algorithm that can identify new coronavirus diseases through cough sounds is “a classic example of artificial intelligence.”
He said it is like transmitting a large number of X-ray images to a machine, and eventually the machine can learn to identify cancer cases from the X-ray images.
Chase is focused on studying the possible impact of artificial intelligence on humans and society in the future, and has also warned about the impact of artificial intelligence.
But this time Chase said, “This is an example of using artificial intelligence to help us. From this point, I don’t see any danger in artificial intelligence.”
