Masks are an important means to contain the novel coronavirus and prevent the spread of an infected person via the respiratory tract. Many states rely on state-regulated obligations. However, such a legal obligation is also accompanied by a test to prevent people without masks from crowds.
Thanks to advances in both theoretical AI research and practical, technical implementation, such tests can be carried out very cost-effectively by machine. However, there is a danger of mass monitoring without cause if too much data is collected.
In order to ensure a compromise between technical convenience and the protection of basic rights, Niklas Kühl, Dominik Martin, Clemens Wolff and Melanie Volkamer are investigating how mask recognition can be implemented in a way that protects the private sphere. This project has now resulted in a publication that can be viewed in the KIT library.