05. – 10. Juni 2005, Dagstuhl Seminar 05231
Scientific Visualization: Challenges for the Future
Auskunft zu diesem Dagstuhl Seminar erteilt
Pressemitteilung vom 24.05.2005: "Wenn Computerbilder Unsichtbares sichtbar machen" (German only)
Scientific visualization (SV) is concerned with the use of computer-generated images to aid the understanding, analysis and manipulation of data. Since its beginning in the early 90's, the techniques of SV have aided scientists, engineers, medical practioneers, and others in the study of a wide variety of data sets including, for example, high performance computing simulations, measured data from scanners ( CAT, MR, confocal microscopy ), internet traffic, and financial records. Somewhat as a result of these past successes, matters are changing for research in SV. The data sets are becoming massive in size, complex and multi-dimensional in nature and the goals and objectives of the visualization much less precisely defined, but yet the results are needed with higher urgency and importance. The multiresolution and hierarchical methods of today do not scale to these new data sets. The segmentation and knowledge extraction methods of today need to be completely revamped in order to be useful. Because of the changes that are taking place in SV, it is important that a group of senior researchers meet with select junior researchers to map out the future research agenda for this critical area.
Specific Themes of the Seminar:
- Ubiquitous Visualization.
- As ubiquitous computing is getting increased attention, also visual display of everywhere available data is necessary. Challenges include: heterogeneous output devices, novel interaction metaphors, network bandwidth (availability, reliability), graceful degradation of algorithms with respect to largely varying resources, invivo visualization (real time, no pre-processing, robust).
- Categorical Visualization.
- Information and knowledge is extremely difficult to extract from multi-valued, multi-dimensional, multi-modal and multi-layered categorical data. These data sets abound today and the pay-offs for understanding them are substantial. Mathematical techniques based upon functional relationships break down requiring completely new paradigms to visualize these types of data sets.
- Intelligent/Automatic Visualization.
- Ever-increasing data sizes require semi-automatic methods that concentrate on the typically very small portion of the relevant information in the data. Techniques include model- and knowledge-based segmentation, classification in abstract feature spaces, computation of saliency information from derived data characteristics, automatic detection of important isosurfaces, automatic creation of expressive transfer functions, automatic landmark selection and automatic path and navigation guidance.
- Point-based/Mesh-free Visualization.
- A typical strategy to visualize unorganized multidimensional data sets is to transform the data into standard geometric primitives of triangles and triangular mesh surfaces prior to rendering. This intermediate step is time consuming, but necessary to map the data set to standard (hardware and software) graphics primitives. With the recent advances in point-based rendering, new efficient and creative approach for visualizing scattered and unorganized data sets are potentially possible.
Dagstuhl Seminar Series
- 18041: "Foundations of Data Visualization" (2018)
- 14231: "Scientific Visualization" (2014)
- 11231: "Scientific Visualization" (2011)
- 09251: "Scientific Visualization" (2009)
- 07291: "Scientific Visualization " (2007)
- 03231: "Scientific Visualization: Extracting Information and Knowledge from Scientific Data Sets" (2003)
- 00211: "Scientific Visualization" (2000)
- 9724: "Scientific Visualization" (1997)
- 9421: "Scientific Visualization" (1994)
- 9135: "Scientific Visualization" (1991)