Figure 6. Visualization generated for the first option search task. The goal selection is encoded as a height field. Note that if AutoVisual had selected a point cloud as the encoding technique, the hierarchy would have only two levels.