Contextual Grouping of Labels

Learning materials frequently employ illustrations with many labels in order to coordinate visual and textual elements. The labels in illustrations support mainly two search tasks: learners can determine the graphical reference objects of unknown terms or vice versa (i.e. get textual descriptions for unknown visual objects). In traditional print media, human illustrators often overload an illustration with many labels in order to reduce the printing costs. But such illustrations ignore the limitations of human cognitive processes. Based on the chunking principle, we argue, that interactive 3D visualizations should be carefully adopted to the current search tasks. Therefore, we consider semantical relations to select only those labels which are relevant for the current task and emphasize the associated graphical objects. Finally, we present a novel layout algorithm for label grouping to aid learning.