Generic model-based application of modular image processing chains for medical 3D data analysis in clinical research and radiographer training

During the last two decades, radiology underwent with the introduction of digital diagnostic technologies a revolution. But these novel technologies are still only sparely manifested in the curriculum of radiographers with respect to image processing and practical training. In this work we introduce a process model for applying image processing modules in a generic way also suited for students and medical staff without deep knowledge in computer sciences. The systematic use and parameterization of different filtering strategies, region growing and life wire technology provided by MeVisLab rapid prototyping application allows for semi-automated segmentation and analysis of diverse anatomical structures for medical diagnostics. In this way future radiology staff gains insight into segmentation tasks regularly achieved by highly specialized software or research prototypes. Besides applicability to the educational domain, the methodical image processing chain is perfectly suited for generating unitized reference segmentations in clinical research for testing and validation of image processing research prototypes.