Other Free Encyclopedias » Online Encyclopedia » Encyclopedia - Featured Articles » Contributed Topics from P-T

Scientific Visualization - Introduction, What is?, Visual Representation in, for Other Fields

data rendering information images

The University of Sydney


Visualization is the process of transforming data into some sort of graphical representation. The data visualized may be real numbers, abstract theoretical quantities or relationships, or may reflect a series, a gradation, or a change in some quantity with respect to others. These datasets are often thought of as image information. The visualization process maps them onto their inherent variables, which determine the shapes and appearances of pictorial figures representing the data. These pictorial figures are then placed on a display screen. The main task of these figures is to convey as much information about the dataset as possible when they are displayed and observed by viewers. Because the eye and brain are intimately involved in this process, visualization requires some understanding of the workings of the human visual system and perception. Visualization techniques have also been widely adopted by the entertainment industry.

There are two different types of visualization depending on the types of data they are concerned with. Information visualization usually deals with abstract data such as sociodemographic, linguistic, and financial. These types of datasets usually do not have physical shapes and appearances associated with them. Scientific visualization, on the other hand, aims to construct images of scientific and engineering data obtained through experiments on, and simulation of, physical subjects or theoretical concepts such as the interiors of stars or the accretion disk of a black hole. Hence, they usually have physical geometry or appearances associated with them.

What is Scientific Visualization?

Scientific visualization was officially recognized in the Report of the National Science Foundation’s (NSF) Advisory Panel on Graphics, Image Processing, and Workstations published in 1987. It was introduced as “visualization in scientific computing” and was described as a discipline covering data representation, data processing, user interfaces, and visual representation including multi-modal sensory representation. The necessity of scientific visualization emerged as a result of rapid advances in computing and electrical engineering technologies, especially high-performance computing during the mid-1980s. Since that time, scientists and engineers have been flooded with the increasing amount of data from experimental equipment and computer simulators. They are now required to process these massive and complex datasets to understand the various natural phenomena they have observed and to evaluate validities of theoretical models. To assist the process of understanding complex natural phenomena, a suite of technologies, which compose scientific visualization, needed development.

The basic elements of scientific visualization are listed below.

Data acquisition through experiments and simulations — Data for scientific visualization can be obtained through various numerical analyses such as a finite element method, and experimental and other observational processes such as satellite imaging and medical imaging. Data processing to transform, extract, and enhance information — If the raw data is thought to be unsuitable for deriving appropriate geometries and visual appearances, the data would require adequate data transformation processes. Such processes would enhance and possibly extract information in different forms. Computer graphics — Two- and three-dimensional geometrical modeling, rendering, and animation processes are used to convert scientific data to a displayable form. Observation and interaction — Visualization is an iterative process and a user may generate new data through interactive exploration of visualized data to gain better understanding of the nature of the information.

Visual Representation in Scientific Visualization

Among the four elements described above, visual representation derived through modeling and rendering processes may have the most significant impact on how visualized images are perceived by viewers. In particular, good selection of rendering methods and parameters could produce striking photorealistic images. Various global illumination algorithms are usually used to create such photorealistic renderings.

There are a number of global illumination algorithms regularly used in scientific visualization such as ray tracing, radiosity, and photon mapping. Ray tracing, for instance, is a method of tracing the path (ray) of the light from a light source to a surface of an object. Although it can generate very good realistic results, it is a computationally heavy process. To overcome this problem many algorithms have been developed. A radiosity algorithm computes effects on surfaces caused by photons from light sources. Such effects are used by a recursive ray tracing algorithm to create a more realistic image.

Photon mapping is based on a ray tracing algorithm and simulates the interaction between light and different objects. This algorithm is particularly useful in computing the refraction of light through transparent materials such as water, and the scattering by particulate materials like smoke.

Designers and programmers of scientific visualization systems now have to consider the influence of the human visual system on how such visual appearances are perceived. Along with the parameters of rendering, colors and parameters of animations are also important in effectively conveying information. How such parameters would affect human perception is studied more rigorously in the field of information visualization. This is because it typically deals with abstract data, which are often unassociated with any physical shapes and appearances. Hence, the selection of appropriate geometries and visual features for rendering holds the key to generating good information visualization.

Scientific Visualization for Other Fields

Scientific visualization as a discipline may have started as a way to assist scientists and engineers to improve and strengthen their understanding of their subjects of study. However, many software and hardware technologies developed for scientific visualization are now routinely used in other fields and industries.

Many medical imaging systems, such as ultrasound imaging, X-ray, computed tomography (CT), and magnetic resonance imaging (MRI) are non-invasive and produce various types of images revealing the internal anatomies of samples. These medical imaging systems typically produce cross sections (image slices) of a sample. Although those sliced images contain an enormous amount of useful information, only trained medical practitioners are normally able to recognize it. Scientific visualization, especially volumetric visualization in this case, is actively used to produce a three-dimensional image from a set of sliced images. Such volumetric representation is not only helpful for interpreting and improving understanding of otherwise unobtainable medical images, but it is also useful for training purposes.

Another industry that fully utilizes the technologies developed for scientific visualization is the entertainment industry. Movies and television programs such as Shrek, Matrix , and Walking with Dinosaurs have used 3D geometrical modeling techniques to create imaginary creatures, dinosaurs, worlds, and environments of the past. Although both physical modeling (such as clay and wire frame modeling) and computer modeling consumes a large amount of the film and television production process, models of those objects themselves do not give an impression of reality, especially when animated. The realistic representations seen in those movies and television programs are the result of various rendering techniques. Since the early 1990s many photorealistic rendering systems have been developed. Some of them were developed specifically for the entertainment industry and fully exploit the state-of-the-art ray tracing methods, radiosity methods, and global illumination models. Pixar’s RenderMan is probably the most well-known rendering system among other similar systems (such as YafRay and Mental Ray) because of its high rendering quality. RenderMan is an industry standard and has many features such as shading language, anti-aliasing, and motion blur to produce photorealistic renderers. Such rendering systems are usually used to produce the final output images for 3D modeling systems like Maya, Houdini, and Blender.

Scientific visualization has gone through significant changes and improvements in the last couple of decades, supported by marked advances in computing and graphics hardware technologies. Modern commodity computers and graphics cards are now powerful enough to carry out many sophisticated scientific visualization tasks. Scientific visualization is no longer just for “making the invisible visible.” It now allows for the invisible imaginings of both scientist and artist with very realistic appearances due to the recent advances in rendering techniques. Such eye-catching and striking photorealistic images are very effective in appealing to human visual systems and are able to convey information previously hidden in the invisible non-pictorial datasets. Although many modeling and rendering techniques were developed for scientific visualization, they are now used for other fields such as the entertainment industry. Furthermore, many computer graphics techniques developed for non-scientific fields are now used to advance scientific visualization. It is getting much more difficult to define a clear boundary between scientific visualization and artistic computer graphics, though the motivational differences, scientific understanding, and entertainment remain distinct.

[back] Science as Art

User Comments

Your email address will be altered so spam harvesting bots can't read it easily.
Hide my email completely instead?

Cancel or