The cytoskeleton is a highly dynamic intracellular network of protein fibers that provides the cell with stability, plasticity and motility.
In this work, methods for the analysis of confocal laser scanning fluorescence microscopic images and image series are introduced, designed to reveal the structural and dynamic properties of the keratin cytoskeleton in living cells.
To deal with the low signal-to-noise ratio of live cell confocal fluorescence images, a model for signal and noise transfer in confocal laser scanning fluorescence microscopy is introduced. It enables effective noise reduction as well as simulation of the imaging process. Thus, synthetic data is generated and used throughout the work for the quantitative evaluation of the image analysis methods. Methods for the extraction of the network structure by 3D segmentation as well as for the analysis of the 3D filament motion and turnover are presented. In particular, keratin turnover analysis reveals the amount and localization of filament polymerization.
Application of the proposed methods in large-scale experiments demonstrates their use for cell biology research.