COVID-19 self-tests have been an integral part of many people's everyday lives for some time now. Even blind and visually impaired people can perform the tests independently, but they often rely on the assistance of others when reading the results. Since the beginning of the year, employees of the Innovation Lab in the IT Department have been working on a concept for the automated recognition of COVID-19 self-tests for blind and visually impaired people. Using artificial intelligence, software will evaluate photos of rapid COVID-19 tests and determine the correct result. With this concept, the team has now won the "TUM.ai Makeathon AI4SocialGood."
The Makeathon was hosted by the Technical University of Munich. More than 250 students, working in teams, had 48 hours to develop solutions for one of the challenges, program a live demo version, and present their results to the jury. The challenges were submitted by nine companies and NGOs on the topic of "Artificial Intelligence for the Common Good."
Covision “has the potential to make the lives of the approximately 100,000 visually impaired and blind people in Bavaria a little better.”
The IT department team called on city employees to submit photos of their rapid tests to feed the artificial intelligence analysis system with data. At the initiative of the IT department staff, the artificial intelligence-based app solution Covision was designed at the Makeathon. The concept impressed the jury, and the Covision team was named the overall winner. The prototype is already available online; the exact date for implementation of the winning concept is currently unknown.
Steffen Erzgraber, State Director of the Bavarian Association of the Blind, says:
"COVID test detection has the potential to make the lives of the approximately 100,000 visually impaired and blind people in Bavaria a little better and to ease their situation, especially during the pandemic. I'm very curious about the further development of the Covision app!"
