As datasets usually contain large regions which are classified as transparent, several methods have been suggested to rapidly traverse empty space. Levoy presented an approach called hierarchical spatial enumeration [25]. The algorithm first creates a binary pyramid of the volume, which encodes empty and non-empty space. Raycasting is started at the top level of the pyramid. Whenever a ray reaches a non-empty cell, the algorithm moves down one level, entering whichever cell encloses the current location. Otherwise, the intersection point with the next cell is calculated and the ray is forwarded to this position. Following this idea, a min-max octree based on the volume's data values can be generated. This octree can be used to efficiently create the pyramid data structure whenever the classification changes. Another approach is space leaping [4,50,6]. Here, a distance transform is applied to the volume to calculate a "proximity" or "skip" value for each empty cell which encodes the distance to the nearest opaque cell. The value therefore is the distance that can be safely skipped along any ray that samples this cell. A drawback of this method is that it requires extensive processing every time the transfer function is changed.