On automation of brain CT image analysis

During recent years a number of medical diagnosis support systems have been presented. Such systems are important from medical point of view, because they aid physicians in their daily work. Some of those systems, like Computed tomography support systems (CT) rely on image data analysis, making it an interesting research topic from pattern recognition point of view. The paper presents a practical realization of a medical support system, including segmentation, feature extraction and decision processes. Various techniques are used within the system: fuzzy logic, shape descriptors, classifiers and automatic image annotation framework. Series of 2D CT images are an input of the system. As a result, a patientpsilas diagnosis is presented.