2023-04-17 21:57:39
It might be the end of the road for employees who try to miss work by claiming a cold. An artificial intelligence program achieved promising results in identifying whether people were really sick just by electronically analyzing the voices of volunteers.
The program achieved 69.9% accuracy in the diagnoses. The index is still too low to be used in offices, but the research was able to establish parameters of voice frequency that can identify who is sick.
To arrive at this result, the test consisted of making people count from 1 to 40 and read a short children’s story by Aesop. With this, the program identified the moments when the voice presented unexpected alterations, characteristic of those who have a runny nose or sore throat.
Methodology
The research in question was developed by programmers from the Indian University of Sardar Vallabhbhai and published in Biomedical Signal Processing and Control. For the study, excerpts from the voices of 630 Germans were analyzed, who subsequently underwent blood tests to prove their diagnoses.
The test, according to the researchers, was created to reduce the number of people who circulate with the symptoms and end up infecting other individuals. “Detecting these symptoms may be important to prevent the spread of viral infections by allowing you to monitor the health of patients remotely,” say the researchers in the study.
0
Although the conclusions indicate that the parameter “efficiently captured the different voice modulations of the cold ones”, revisions will still be made to improve the program’s performance.
Get news from metropolises on your Telegram and stay on top of everything! Just access the channel: https://t.me/metropolesurgente.
Have you read all the Health notes and reports today? Click here.
1681771889
#Robot #identifies #colds #voice #study