Entwicklung eines halbautomatischen Programms zur bildanalytischen Auswertung der Morphologie von Neuronen mittels Sholl-Analyse

Kurzfassung

This thesis deals with the development of a program for the semi-automatic analysis of (fluorescence) microscopic images. Sholl analysis is used to investigate the morphological complexity of neurons that are altered by neurotoxic influences such as certain agents or radiation. The latter is of importance in the risk assessment of interplanetary travel of humans. The developed application has a complete graphical user interface and the possibility to document the analysed images with data from the experiments. For image processing, the OpenCV libraries were used. The images are first segmented. The origins of the outgrowing neurites can be determined by a developed method. Originating from these somatic starting points they are traced with a ChainCodebased algorithm. The implemented Sholl analysis uses mathematical formulas to determine the number of intersections between neurites and a circle whose centre lies in the soma of the cell. The radii and step sizes required for this can be given in pixels or µm. To simplify the conversion, CZI files can be loaded from which the corresponding metadata can be imported. Each intermediate step can be saved as an image and the analysis results are exported as a CSV file.