Projekt Beschreibung

White Matter Hyperintensities (WMH), lacunar stroke lesions and incident infarcts are MRI biomarkers which indicate cerebral microangiopathy, a risk factor for stroke and vascular dementia. In collaboration with our partners at the Frauenhofer Institute for Digital Medicine MEVIS and mediri Medical Imaging Research Institute, we develop automated imaging segmentation and quantification tools to apply in large-scale imaging datasets from a population at risk for cerebrovascular disease. We have received funding by German Federal Ministry for Economic Affairs and Energy for our current project DELIGHT (Deep Learning for quantification of hyperintense brain lesions). In DELIGHT, we train Deep Neural Networks (DNN) to segment WMH, based on data pooling from various, heterogeneous data sets of stroke patients and patients with presence or increased risk for cerebral small vessel disease.