Purpose and goal
The MIDAS project aims at solving the problem of anonymity when collecting video data in real traffic environments. MIDAS will develop ML algorithms for replacing faces, license plates and any other recognizable marks with alternative fake instances of those same objects. Once the important or sensitive information has been replaced, video data can be saved for future use while complying with the GDPR guidelines.
Expected results and effects
The project will develop methods that make it possible to save video data in an anonymized way that allows you to save video data without compromising GDPR. The results will facilitate data management during training and development of ML-based algorithms, which are basic technology in, among other things, active safety and automated driving.
Planned approach and implementation
The project is divided into four work packages, project management, data collection, anonymization and demonstrator. Project management and demonstrator is led by RISE, data collection is led by Viscando and anonymization is led by Berge. Berge will hire an industrial doctoral student who is supervised by Halmstad University.