Semantic Depth of Field – Using Blur for Focus+Context Visualization

One central task in information visualization and related fields (like volume and flow visualization) is displaying information in a context that makes it easier for users to understand. Blur is a visual cue that has been playing an important role in photography for over 150 years, but has been widely ignored in computer graphics. Sharp objects in photographs immediately attract the viewer’s gaze – distinguishing between sharp and blurred objects therefore is very well suited for directing the viewer’s attention to certain objects or parts of
the image. In this thesis, a method called Semantic Depth of Field (SDOF) is proposed, which blurs
currently irrelevant objects and thus guides the viewer’s attention. This method only requires one additional value for each data point: its relevance. This relevance is then trans- lated into a blur level, which is used for drawing objects. The user has full control over the functions involved in this process. A number of applications is shown to demonstrate the usefulness of SDOF.
Because blur is known to be slow, a method for fast blurring of objects is also presented, which makes it possible to use SDOF in interactive applications.
The results of a user study are also presented, which showed that SDOF is a preattentive feature, i.e., can be perceived within 200 ms, and does not require serial search. SDOF is also not significantly slower than color when used in search tasks; and it does not decrease performance when combined with another feature, as is usually the case.
Stacks Image 150
Robert Kosara,
Semantic Depth of Field – Using Blur for Focus+Context Visualization,
PhD thesis, Vienna University of Technology, Austria, 2001.