Scientists design a method based on artificial intelligence to detect Covid from a patient's voice or cough

The technique, which has been developed by researchers in Granada, Murcia and France, has proven effective in 77 per cent of cases so far


A team of international scientists from Granada university, the Eurecom university in France and engineers at the Biometric Vox company in Murcia has designed a method based on artificial intelligence to detect Covid-19 from voice and cough recordings. So far, it has proven effective in 77 per cent of cases.

The researchers developed a technique which automatically analyses the recordings and looks for sound patterns which could indicate that the person has the virus.

“As we know, among other symptoms, patients with Covid have a dry cough, shortness of breath and faster respiration, so the method checks for alterations in the patient’s voice which suggest they have the illness,” explained José Andrés González López, a professor in the Department of Signal Theory, Telematics and Communications at Granada university.

The advantage this method has over others is that the technology is not invasive and is instant, so it could be used via an App or phone call as a screening or rapid diagnosis method for Covid-19, without the patient having to go to the health centre and put other people at risk if they are positive.

The researchers also said that this voice recognition technology, which can be applied to coughs as well, can also be used in the future to detect and monitor similar illnesses which affect the respiratory tract, such as flu and colds.