Abstract
This paper presents a segmentation pipeline for computer-based automatic analysis of multimodal tomographic images. It is a computer based support for the localization of pathological tissues such as brain tumors. The segmentation pipeline of the presented approach includes texture analtsis, classification with a modified Kohonen Feature Map, a collection of classifiers and knowledge based morphological postprocessing. Furthermore this paper presents a statistical investigation that compares the wavelet transform to classical texture analysis methods. Patient data which was acquired using magnetic resonance imaging (MRI) and computer tomography (CT) is used for this investigation.