Artificial intelligence technology or AI plays a role in the diagnosis of various diseases. More recent developments in the medical field during the COVID-19 pandemic. Research team in the US Artificial intelligence development In the detection of superwallet COVID through coughing
Believe that many people When you hear a cough One of the symptoms of the flu Including COVID-19 Causing tremors to tremble together at the same time But now the US research team is bringing machine learning systems to help classify a person's voice patterns. Indicating COVID-19
Importantly, this system detects COVID by voice. Highly accurate in detecting COVID-19 In infected people who do not show clear symptoms about this epidemic. And it could be an important means of fighting the COVID-19 pandemic. That has infected nearly 50 million people and died more than 1 million people around the world now.
A research team at the Massachusetts Institute of Technology, or MIT, has released reports of testing artificial intelligence systems, or AI, to examine people's voices. The sound of the vapor throttle through computers and mobile phones. Including information about various medical conditions Of those people It contains information on people with a common cough, sick cough, and coughing from COVID-19.
The MIT research team has obtained more than 70,000 audio recordings, including 200,000 human samples, to provide a database for AI systems to process the vapor. Including the normal speech of the person
Analysis Development Team "AI covid" found that the AI system was able to accurately analyze 98.5% that the cough came from people confirmed to have COVID-19. And in the same system Can detect COVID-19 Through the coughing of people with COVID-19 Asymptomatic, 100% accurate
Brian Subirana, a scientist at MIT's Auto-ID Laboratory, the lead researcher, revealed that the team developed an AI model to analyze the vapor throttle sound. To look for signs of Alzheimer's disease has come. Because these signs include behavioral changes and memory loss. While Alzheimer's disease Causing problems with the brain and muscles as well This includes a diminished ability to speak. And found that the detection of Alzheimer's disease through coughing It is one breakthrough in detecting signs of this disease.
After the coronavirus pandemic has spread around the world, Subirana sees the disease analysis model through sound coughing. May work with COVID-19 And in testing it was found that Analysis model for Alzheimer's disease and COVID-19 Are similar Changes in the voices of people infected with COVID-19 Which can be measured immediately even without showing any other symptoms
Research team aims to develop application to measure COVID-19 That is easier to use Just load the app The users' coughs are recorded on their phones and sent back for preliminary analysis if COVID-19 is suspected. Before going to be tested for infection later
The head of the study stressed that this app could be useful in reducing the spread of COVID-19 if everyone turned to this initial measurement. Before going to study in class Before returning to work in the factory Or go to a restaurant, etc.
In addition to mitochondrial analysis of MIT, in the US there is a Carnegie Mellon University research team. In Pennsylvania, machine learning is used to develop a voice-activated COVID-19 detection system that uses voice recording of the vapor. Pronunciation of vowels and consonants To identify signs of coronavirus infection too As reported by the Pittsburgh Post-Gazette.
And that England has a project like this Prepared by the University of Cambridge. Which the research team in England It was revealed in July that it has developed a machine learning system that can identify COVID-19 infection. Through a person's coughing and breathing sounds Which is about 80% accurate