The availability of cellular markers tagged with the green fluorescent protein (GFP) has recently allowed a large number of cell biological studies to be carried out in live cells, thereby addressing the dynamic organization of cellular structures. Typically, microscopes capable of video recording are used to generate time-resolved data sets. Dynamic imaging data are complex and often difficult to interpret by pure visual inspection. Therefore, specialized image processing methods for object detection, motion estimation, visualization, and quantitation are required. In this review, we discuss concepts for automated analysis of multidimensional image data from live cell microscopy and their application to the dynamics of cell nuclear subcompartments.