Image Processing Reference
In-Depth Information
lead to interesting visualization research. Highly interesting and very challenging,
also, the emergence of higher-dimensional scientific data (in the sense of a higher-
dimensional domain) leads to new visualization questions. Multi-run/ensemble sim-
ulation data, for example, includes parameters as additional independent variables.
New approaches are needed to deal with this situation, especially in the context of
scientific visualization, where generally a stronger and more immediate relation is
present between the domain of the data and the visualization space (and to estab-
lish this relation in an effective way becomes more challenging, obviously, the more
dimensions the data domain has). The integration of descriptive statistics, for exam-
ple, is one opportunity that allows to perform a linked interactive visual analysis
both on aggregation level as well as on the original multi-run data. It seems clear,
however, that more research is needed to more thoroughly discuss, what the best
possible approaches are.
Robert S Laramee on Spatial Integration:
Another major challenge of multi-field visualization is the integration (or coupling)
of two or more data fields into the same spatial domain from which they originate. A
common example is from computational fluid dynamics (CFD) [
9
]. CFD simulation
data generally contains many attributes, e.g., flow velocity, pressure, temperature,
kinetic energy, etc. And each multi-attribute data sample is associated with the same
spatial domain. It is tempting to separate each attribute into its own visualization
space, either abstract or scientific. However, integration of the data attributes into
the same spatial domain from which they stem offers distinct advantages. However,
how can such an integration be done in a meaningful and helpful way without over-
crowding the visualization space?
Lars Linsen on Intuitive Visual Exploration of Multi-variate Features:
Features may have a complicated geometrical structure in the multi-dimensional
attribute space. Extracting those features interactively is often tedious, if not impossi-
ble. Automatic components can help to compute such features. However, an intuitive
visual exploration of such features is crucial to the user's understanding. What is the
object space representation and, more importantly, what attribute values correspond
to such a feature? Are their other features that are related, which possibly should
have been merged by the automatic component? How homogeneous is a feature?
Are their sub-regions within a feature that allow for further splitting of the feature?
Such questions shall a user be able to answer when exploring the multi-field data.
Intuitive visual encodings in object- and attribute-space as well as intuitive interaction
mechanisms need to be provided.
Klaus Mueller on Channel Fusion:
The term “channel” is often used in the context of color images, comprised of a reg-
ular array of RGB color pixels. By mapping these 3D vector data to the three display
primaries, channel fusion can occur directly in the viewer's visual system, engaging
the tristimulus processes of color perception. However, once the number of channels
exceeds three, the fusion must be externalized via some analysis and subsequent
Search WWH ::
Custom Search