Research project

Whole body magnetic resonance imaging (MRI) combines the advantages of high tissue contrast with the possibility of scanning the whole body of patients without the need of repositioning. However, due to the large amount of acquired data, a prompt and reliable screening is a challenging task for radiologists. Due to the heterogeneity of the data sets, an automatic analysis of the images is very difficult. Furthermore, a missing gray value standard like Hounsfield units in computed tomography makes prior knowledge about gray levels and their distributions infeasible, but many segmentation and classification approaches rely on this information.

The project deals with the development of combined methods for inter and intra intensity variation correction using statistical information of the data sets in order to create an protocol dependent intensity standard for MRI. Consequently automatic segmentation and classification algorithms can be used much more easily in MRI.