Anchored in Rostock, connected across Europe, from document AI to medical and maritime research.
Explainable-AI radiology assistant for dementia diagnostics from MRI.
DEJ Technology · DZNE · Universitätsmedizin Rostock (IfDIR)
KI-DERA develops and validates a radiological AI assistance system based on explainable AI that supports dementia diagnostics and differential diagnosis from MRI images. The project runs for three years with DEJ Technology GmbH, the German Center for Neurodegenerative Diseases (DZNE) and the University Medical Center Rostock (IfDIR).
LLMs and multi-agent architectures for privacy-compliant, explainable document workflows.
ALTOW Digital Innovation · Universität Rostock (CITlab)
SPOC-AI builds an AI-based assistance system on language models and multi-agent architectures that supports complex document-based workflows in a privacy-compliant and explainable way. The project runs for three years with ALTOW Digital Innovation and the University of Rostock (CITlab).
Generic AI foundation model for object recognition in the maritime underwater domain.
EvoLogics · GEO-DV · Subsea Europe · IOW · Uni Rostock · Fraunhofer IGD
OTC-DaTA2Model develops a generic AI foundation model for underwater object recognition and a modular architecture for probabilistic digital twins. These combine data from different sensor systems with background knowledge to estimate the current state and future development of marine environments. The project runs for three years with EvoLogics GmbH, GEO-DV GmbH, Subsea Europe Services GmbH, the Leibniz Institute for Baltic Sea Research Warnemünde (IOW), the University of Rostock and Fraunhofer IGD.
Genomics and marine environment analysis.
Leibniz IOW · LGC Genomics · HYDRO-BIOS
OTC-Genomics2 builds on the results of OTC-Genomics: the molecular monitoring toolbox for aquatic habitats is extended with metagenomic analyses and advanced AI methods and tested in the regular environmental monitoring of LUNG and BSH. The project runs for three years with the Leibniz Institute for Baltic Sea Research Warnemünde (IOW), LGC Genomics GmbH and HYDRO-BIOS Apparatebau GmbH.
Medical image analysis.
PolyDiagnost · Universitätsmedizin Göttingen
CAPTAIN uses artificial intelligence to annotate multimodal endoscopy images in pancreatic cancer in real time, so tumor cells can be detected during the examination and treated or removed directly. The project runs for three years with PolyDiagnost GmbH, the Institute for Diagnostic and Interventional Radiology at the University Medical Center Göttingen and the Faculty of Engineering & Health of the University of Applied Sciences and Arts.
Maritime AI and probabilistic digital twins from heterogeneous sensor data.
EvoLogics · IAV · Fraunhofer IGD · Uni Rostock · IOW
The Ocean Technology Campus project DaTA looks at data from several contexts as a whole and evaluates it in one consistent way, for example payload and control data that have so far mostly been collected and processed separately. The project runs for three years with EvoLogics GmbH, IAV GmbH, Fraunhofer IGD, the University of Rostock and the IOW.
Genomics for marine environmental monitoring.
Leibniz IOW · LGC Genomics · Hydrobios · Fraunhofer IGD
The Ocean Technology Campus project Genomics extends existing environmental monitoring of aquatic habitats with new analyses based on microbial nucleic acids (16S rRNA genes) and environmental nucleic acids (eDNA; 18S rRNA genes) from water samples. The project runs for three years with the Leibniz Institute for Baltic Sea Research Warnemünde (IOW), LGC Genomics, Hydrobios and Fraunhofer IGD.
EU-funded radiological assistance system.
Universitätsmedizin
The Intelligent Radiological Assistant evaluates imaging such as X-ray, CT and MRI, with a focus on thorax scans and reasonable, explainable AI. The project runs for three years with the University of Rostock.
Recognition and layout analysis for historical newspapers and handwriting (EU).
Uni Rostock · La Rochelle · Helsinki · Innsbruck · Montpellier · Wien · Österr. Nationalbibliothek · Nationalbibliothek Frankreich
NewsEye makes historical texts understandable and extracts information from them, such as names, places and other details. The project runs for three years with the University of Rostock, the University of La Rochelle, the Austrian National Library, the University of Helsinki, the University of Innsbruck, the National Library of France, the University of Montpellier and the University of Vienna.
Recognition of historical handwriting from European heritage, 1500 to 1800 (EU Horizon 2020, grant 674943).
Uni Rostock · Greifswald · Nationalarchive FI/NO/SE · Erlangen-Nürnberg · Innsbruck · Valencia · Edinburgh · Wien
READ recognises historical handwritten texts from European cultural heritage between 1500 and 1800. The project runs for three years with the University of Rostock, the University of Greifswald, the National Archives of Finland, the University of Erlangen-Nuremberg, the University of Innsbruck, the University of Valencia, the University of Edinburgh, the National Archives of Norway, the Swedish National Archives and the University of Vienna.
Automatic full-text recognition in historical handwritten documents (EU).
EU-Konsortium
The tranScriptorium project develops algorithms for automatic full-text recognition of handwritten historical documents. The project runs for three years with the University of Rostock.
Self-organizing recurrent neural networks for language processing.
Uni Graz · INSERM · Uni Ghent · Jacobs University Bremen
ORGANIC researches self-organising recurrent neural networks for speech processing. The project runs for three years with the University of Graz, INSERM, Ghent University and Jacobs University Bremen.
