Detection and delineation of pathology, such as cancerous tissue, within Multispectral brain magnetic resonance (MR) volume sequences is an important problem in medical image analysis. For example, a precise and reliable segmentation of brain tumors present in the childlike brain is regarded critical when aiming for the automatic extraction of diagnostically relevant quantitative or more abstract findings. This may include the volume of the tumor or its relative location. Once these findings are obtained they can be used both for guiding computer-aided diagnosis and therapy planning as well as for traditional decision making. The goal of this project is the development of a fully automated approach to segmentation of pediatric brain tumors in multi-spectral 3-D magnetic resonance images.