Projects and Products

Generative AI models enable the creation of realistic images with great potential for creative and medical applications. At the same time, they pose risks, such as through their targeted use for disinformation. The RealOrRender project aims to develop methods that can reliably and robustly distinguish AI-generated images from real photographs.

For the “Intelligent Video Surveillance“ project, Fraunhofer IOSB's VID department is developing a system that can detect possible crimes based on movement patterns and alert police officers in the command and situation center. Facial recognition is not part of the solution approach.

Understanding real traffic - safeguarding automated mobility. The safe introduction of highly and fully automated vehicles depends crucially on how well we can digitally map today's traffic events - especially critical situations. AVEAS develops scalable, data protection-compliant methods for systematically recording and modeling risk-relevant scenarios and transferring them to virtual test environments.

Multisensory system for AI-based drone detection and classification with situation visualization.

“AutoInspect” is a system for continuous object assessment (in real time) by multimodal inspection along the production cycle. Both a direct reaction to the results and their long-term observation are possible. The system is supported by a digital quality twin with open I4.0 standards.

"S²UCRE" is a project of the Federal Ministry of Education and Research (BMBF). The aim of this project is to improve security in urban environments through crowd monitoring, prediction and decision support.

"ABUL" is an interoperable video evaluation system with integrated real-time and offline procedures and various visualisation components to support the evaluator.

A project of the same name aims to harness our experience across a whole host of AI application domains to realize MLOps with maximum effectiveness and automate it as far as possible. To that end a software suite was developed that incorporates a wide range of tools for creating machine learning applications and training the ML models.

PrognoSF aims to accurately predict local flash floods by analyzing all available data sources with AI-based algorithms.

Economic efficiency, sustainability, ecology and individuality: Intermodal mobility - the use of different modes of transport on one route - has great potential. But the complexity must remain manageable. In other words: Getting to your destination intermodally, for example with a combination of (rental) bike, public transport and e-scooter, must be as easy and pleasant as reaching for your own car key.