Towards Renal Compartment Segmentation Using an Unsupervised Neural Network Approach

¹Computer Assisted Clinical Medicine, Faculty of Medicine Mannheim, University of Heidelberg, Mannheim, Germany; ²Section for Radiology, Department of Surgical Sciences, University of Bergen, Bergen, Norway

ynamic contrast enhanced magnetic resonance imaging is an emerging technique for a more accurate assessment of local renal function. Automated methods mostly involves user interaction or are based on model assumptions.In this work we present a model free and unsupervised approach to renal compartment segmentation in 3D DCE-MRI data. Thereby self organizing maps (SOM)are utilized. Initial results demonstrate that SOMs could be used for a segmentation of the renal compartments but also, could give qualitative insights into local perfusion patterns of the kidney.