The situation experienced with COVID-19 where medical care has been overwhelmed and there is a difficulty to perform rapid and massive tests in an agile way in the population is the starting point of the project developed by CREB UPC (Research Centre for Biomedical Engineering of the Technical University of Catalonia) and funded by the Center for Development Cooperation UPC. The solution seeks to provide a low-cost tool that helps to reduce the burden on the healthcare system and act quickly in case COVID-19 symptoms are detected.

The team led by Prof. Jordi Fonollosa and formed by Guillem Bonilla and Dani Marínhas created a collaborative multi-language tool to easily detect symptoms associated with COVID-19 using the user’s mobile phone. A predictive model exploits the different sounds of cough to predict if the user has symptoms compatible with COVID-19. In particular, the user interacts through the Telegram application installed on their phone and is guided through a series of questions and asked to record a coughing audio. Finally, the predictive model provides information on whether symptoms are compatible with COVID-19.

The system is in the deployment phase for experimental validation. In order to improve the performance of the model, a data collection phase has begun, with special interest in the recordings of people diagnosed with a positive and currently showing symptoms. Anyone can help to improve the model by interacting with the bot and sharing the requested data.

We need you! Help us validate this tool by accessing: https://t.me/Covidscipy_bot