Fraunhofer Institute for Integrated Circuits IIS

2018: Erfolgreiche Evaluierung des Fraunhofer-Anwendungszentrums in Deggendorf, Ausgründung der Abteilung »Computertomographie in der Messtechnik CTMT«, Einrichtung der Forschergruppe »Wissensbasierte Bildverarbeitung« in Passau

Location Passau

Research group for Knowledge-based Image Processing and Visualization (WBV) in Passau

The research group on Knowledge-Based Image Processing and Visualization (WBV) in Passau focuses on the development of three-dimensional image processing and visualization techniques for large data sets, such as those generated in the context of XXL CT or synchrotron CT. Methods of artificial intelligence, classical image processing and visual analytics are used to efficiently process large volume data on standard computers.

Core competencies

Segmentation

An essential step in the analysis of 3D data is segmentation, i.e., the extraction of individual related components from the data sets. The methods developed here enable flexible and interactive segmentation, which can also be used for large data sets and do not require annotated training data.

Compression

Typically, very large data sets cannot be processed on standard hardware. This means that no meaningful interaction with the data can take place. Our method allows a considerable reduction in memory requirements and in addition the extraction of local regions without having to unpack the entire dataset.

Visualization

Our expertise enables us to create customer-specific visualizations that are tailored to the requested use cases and present complex issues intuitively. This provides our customers with a solid basis for their decision-making.

X-ray simulation

Our analytic simulation targets 2D projection generation as well as virtual 3D Computed Tomography. Our application areas are found in the generation of ground truth data for physics-informed AI systems, in optimal scan parameter determination for real scans, spectrum estimation, and finally probability of detection analyses.