The term "visualization for scientific computing," shortened to "scientific visualization," refers to the methodology of quickly and effectively displaying scientific data. Once a computer experiment has been performed, the results must be interpreted. As the volume of data accumulated from computations or from recorded measurements increases, it becomes more important to analyze data quickly. It is also important to design future experiments focused on examining interesting relationships. While looking at statistics can provide insights, it is difficult for the human mind to comprehend large volumes of numbers. Visualization techniques are another approach for interpreting large data sets, providing insights that might be missed by statistical methods.
Scientific visualization is the process of representing raw data output as images that can aid in understanding the meaning of the computer experiment results. However, the purpose of scientific visualization is insight. The pictures provide a vehicle for thinking about the data. Scientific visualization is an external aid to enhance human cognitive abilities for discovery, decision-making, and exploration. Emphasizing parameters of interest using color and highlighting may be necessary to draw attention to particular features. Animation can draw out features dependent on time and help clarify the size and relative position of features.
This technique of converting data into static graphical pictures or dynamic animations works because human beings are inherently better at analyzing information if it is in visual form. The two sides of the human brain function differently. The left side of the brain helps with analytical calculations, verbal communications, and abstract symbols. In contrast, the right side of the brain emphasizes spatial, intuitive, or holistic thinking. It allows the scientist to view an entire complex situation. Graphical representations of data stimulate this part of the brain. Using this approach, a scientist is able to get an overall picture of the data. Later a scientist may use a more analytical method to examine certain anomalies or patterns in the data.
The main reasons for using scientific visualization are the following: it compresses large volumes of data into a limited number of pictures; it can reveal the correlation between different quantities both in space and time; it can furnish new space-like structures beside the ones which are already known from previous calculations; and it opens up the possibility to view the data selectively and interactively in "real time." By following the motions of structures in time, one can gain insight into complicated dynamics. It is also very useful to have the possibility to change parameters interactively and to immediately see the effect of this change. This process is called computational steering. It increases the effective use of central processing unit (CPU) time.
Scientific visualization is used in business to organize, identify, and communicate trends in market and customer data that are obscured or too small to see without amplification. Scientific visualization is used in astronomy to visualize objects known to exist but not directly observed such as black holes, gravitational waves, and collisions of neutron stars. Scientific visualization is used in medicine to visualize cell growth, assist in medical diagnosis, and prepare surgeons for operations. A common example of scientific visualization is the combination of atmospheric data with color, placement, illumination, and texture for use in television weather programs. Two notable scientific visualizations are the visualization of ash plumes from volcanic eruptions, and the visualization of the Visible Man from microtomed traversal and computed tomography scans.
Related Domains and Techniques
Scientific visualization uses techniques from the fields of computer animation, computer graphics, image processing, computer vision, computer-aided design, human-computer interaction, and signal processing. Scientific visualization is closely related to the new field of data mining . While scientific visualization seeks to illuminate known relationships within data, data mining seeks to uncover previously unknown relationships within data.
Images have long been a part of science. In the Renaissance period, Italian astronomer Galileo (1564–1642) interspersed his writings with drawings of the moons of Jupiter and anatomical sketches. Today, more powerful computers and new graphic tools are creating opportunities to visualize events previously too small, too far, or too obscure for a human to see.
Scientific visualization tools range from custom software built specifically for handling certain classes of algorithms to general-purpose visualization systems that support a variety of techniques and handle a wide variety and range of data types. Graphical software tools used in scientific visualization include self-organizing maps, charts, graphs, scatterplots, zoom, sort, overview/detail, fly-throughs, and hierarchies. It is important to differentiate between scientific visualization and presentation graphics. Not all computer graphics packages qualify as visualization systems. Presentation graphics are primarily concerned with the communication of information and results in ways that are easily understood. In scientific visualization, one seeks to better understand the data. However, the two methods are often intertwined.
The two most common underlying approaches to transforming data into visual representations are graphical primitives and data sampling. Graphical primitives involve filtering data to extract information and map these data using geometric primitives such as points, lines, or polygons. Data sampling involves moving data into structured grid fields of different types (uniform, rectilinear, irregular) using interpolation and extrapolation.
Research on New Uses
Scientific visualization has already had a great impact on scientific work by making the evaluation of computer experiments easier and helping the scientist to gain more insight into unknown aspects of scientific problems. Datasets are growing at a substantial rate as researchers explore science in greater detail. Effectively visualizing volumes of data using only one million pixels in the average computer display requires new methods of visualization. Current scientific visualization research focuses on combining data from multiple sources in the same visualization, developing new tools to help users find what they are looking for using visual search engines, and providing collaborative visualization between geographically remote sites.
see also Data Visualization; Pattern Recognition; Physics.
William J. Yurcik
Earnshaw, R. W., and N. Wiseman. An Introductory Guide to Scientific Visualization. New York: Springer Verlag, 1992. Nielson, Gregory M., et. al. Scientific Visualization: Overviews, Methodologies, and Techniques. Los Alamitos, CA: IEEE Press, 1997. Pickover, Clifford A., and Stuart Tewsbury, eds. Frontiers of Scientific Visualization. New York: John Wiley & Sons, 1994.
Annotated Bibliography of Scientific Visualization Web Sites. NASA Ames Research Center. <http://www.nas.nasa.gov/Groups/VisTech/visWeblets.html>