The cytoskeleton is a dynamic scaffold in living cells even in the absence of externally imposed forces. In this study on cytoskeletal deformation, the applicability of hierarchical feature vector matching (HFVM), a new matching method, currently applied in space research and three-dimensional surface reconstruction, was investigated. Stably transfected CHO-K1 cells expressing green fluorescent protein (GFP) coupled to vimentin were used to visualize spontaneous movement of the vimentin cytoskeleton of individual cells using a confocal laser scanning system. We showed that, with proper parameter and configuration settings, HFVM could recognize and trace 60-70% of all image points in artificially translated, rotated, or deformed images. If only points belonging to the cytoskeleton were selected for matching purposes, the percentage of matched points increased to 98%. This high percentage of recognition also could be reached in a time series of images, in which a certain degree of bleaching of the fluorescence over the recording time of 30 min was inevitable. In these images, HFVM allowed the detection as well as the quantification of spontaneous cytoskeletal movements of up to 10% of the cell width. Therefore, HFVM appears to be a reliable method of quantifying dynamic cytoskeletal behavior in living cells.