Protocol

The ability to successfully identify COVID-19 patients from their voices heavily depends on collection of a large dataset that contains speech, coughs and breathing of people diagnosed positive to COVID-19, non-infected individuals, but also people who might suffer from other respiratory conditions.

The data is collected using a multilingual web-based platform which is designed in 8 languages to cover a large range of potential participants.

Besides vocal data, information about the participant’s age, gender, country of residence, native language, weight, height, smoking and drinking habits is also collected, as well as COVID-19 related symptoms and comorbidities.

Health status of the participant (positive or negative to COVID-19) is determined based on the self-declaration confirmed by the standard RT-qPCR or RAT test, with the date of testing.

The participants can also provide the information whether they are currently in home isolation or hospitalized.