Author:
Martin Kraus
Affiliation:
Aalborg University, Denmark
Keyword(s):
Uncertainty visualization, Graph drawing, Information visualization, Volume visualization, Contour tree.
Related
Ontology
Subjects/Areas/Topics:
Abstract Data Visualization
;
Algorithms and Technologies
;
Computer Vision, Visualization and Computer Graphics
;
General Data Visualization
;
Interpretation and Evaluation Methods
;
Presentation Methods
;
Visual Data Analysis and Knowledge Discovery
;
Visual Representation and Interaction
Abstract:
Contour trees can represent the topology of large volume data sets in a relatively compact, discrete data structure. However, the resulting trees often contain many thousands of nodes; thus, many graph drawing techniques fail to produce satisfactory results. Therefore, several visualization methods were proposed recently for the visualization of contour trees. Unfortunately, none of these techniques is able to handle uncertain contour trees although any uncertainty of the volume data inevitably results in partially uncertain contour trees. In this work, we visualize uncertain contour trees by combining the contour trees of two morphologically filtered versions of a volume data set, which represent the range of uncertainty. These two contour trees are combined and visualized within a single image such that a range of potential contour trees is represented by the resulting visualization. Thus, potentially erroneous topological structures are visually distinguished from more certain str
uctures. Moreover, topological structures can be revealed that are otherwise obscured by data errors. We
present and discuss results obtained with a prototypical implementation using well-known volume data sets.
(More)