COVID-19 Sounds

COVID-19 Sounds: AI-Powered COVID-19 Detection

Discover COVID-19 Sounds, an open-source project using machine learning to detect COVID-19 in cough sounds, and learn how it uses AI-powered predictive models for accurate diagnosis.

What is COVID-19 Sounds?

COVID-19 Sounds is an open-source project developed by researchers at MIT to help detect cases of COVID-19 based on analysis of cough sounds. It uses machine learning models trained on thousands of cough sound samples, both from COVID-19 positive patients as well as healthy individuals or those with other respiratory conditions.

The goal of COVID-19 Sounds is to provide a free and easy way for people to screen themselves by simply recording a cough sound sample using their phone and uploading it. The AI model analyzes acoustic features of the cough and predicts the likelihood that it came from someone with a COVID-19 infection.

While it should not be used as a definitive diagnostic tool, COVID-19 Sounds can help identify people who should seek medical testing for COVID-19 based on their cough characteristics. The creators have made the code and models open source to allow others to build on and enhance the accuracy of this approach over time as more cough data becomes available.

COVID-19 Sounds demonstrates the potential of using machine learning on new types of medical data for rapid and low-cost screening of infectious diseases like COVID-19. It shows great promise as a supplementation tool, especially in areas with limited access to swab-based COVID tests.

COVID-19 Sounds Features

Official Links