In this paper, we present some novel animation techniques to help users understand complex structures contained in volumetric data from the medical imaging and scientific simulation fields. Because of the occlusion of 3D objects, these complex structures and their 3D relationships usually cannot be revealed using one image. By our animation techniques, the focus regions, the context regions, and their relationships can be better visualized at the different stages of an animation. We propose an image-centric method which employs layered-depth images (LDIs) to get correct depth cues, and a data-centric method which exploits a novel transfer function fusing technique to guarantee the smooth transition between frames. The experimental results on real volume data demonstrate the advantages of our techniques over traditional image blending and transfer function interpolation methods.