Companies rapidly invested in creating pre-screen applications that predict when anyone has COVID-19 using technology such as speech and smartphone apps. Vocalis Health has developed an AI-based app in a recent creation, which recognizes the spokesmanship of patients with different ailments, like COVID-19, without having to attend a hospital or stressing local health services. The Next Web reports.
It is defined by the company as "a very scalable COVID-19 tool to screen."
In cases where screening processes, such as temperature arms, cannot be used to diagnose asymptotic patients, Vocalis' solution may benefit. Vocalis gathers a lot of voice-related info.
The algorithm of the company extracts 512 features from a voice sample and produces an image - a spectrogram - that can diagnose diseases.
When the firm tries to better diagnose a certain condition, clinical studies first help to determine which kind of speech samples to gather, and which signatures to find in audio.
The pre-training data collection of the organization consists of over 275,000 speakers of different languages to make it globally available.
In order to make COVID-19 more stable and reliable for the future, the company needs to collect more data sets. Moreover, the solution is expected to be sold and made usable for more countries.