Task-based adaptation of graphical content in smart visual interfaces

Visual representation has always been an important communication medium to convey complex facts. In order to be e ective, a visual representation must be adapted to its respective context of use. This is a crucial aspect especially in smart visual interfaces context-sensitive user interfaces that present to the user as speci cally as possible those information and with the level of detail required for her current situation. Of the in uencing factors contributing to this context of use, the characteristics of the content from which the representation is generated and the user's task at hand are the most important ones. The goal of this thesis is to provide a basic approach to the task-based adaptation of visual representations from graphical content in smart visual interfaces, and to detail concrete solution approaches to associated challenges. The objective of the approach thereby is to adapt the graphical content associated with the task at hand automatically in such a way that the content's visual representation provides a good initial view with respect to information relevant to the current working step. In this thesis, two core concepts are introduced that facilitate the automatic generation of task-speci c visual representations from di erent types of graphical content based on a suitable task description. Starting point of the approach is a task model representing a hierarchical decomposition of the task at hand into interrelated subtasks. The page/feature concept represents a method to enrich this model with the speci cation of graphical content associated with individual working steps as well as relevant representation aspects. An adaptation pipeline provides the bridge between the conceptual task context speci cation in the enriched task model and concrete display techniques comprising the functional building blocks of smart visual interfaces. It is discussed how these concepts are applied to graphical content of each of the four principal visual data types raster images, 2D vector and 3D graphics as well as data visualizations. This includes a review on how existing display techniques can be integrated into the adaptation pipeline approach. Moreover, several novel display techniques are introduced that address previously open challenges with respect to task-driven adaptation of visual representations in smart visual interface. The developed concepts and techniques have been practically employed in the scope of a joint industry-academia research project underlining their utility for smart visual interface design.