THE USES OF COMPUTER GRAPHICS

Images can be used for more than just the enhancement of existing material, the addition of pretty pictures to documents. Whilst we are all aware of computer graphics in, say, print and television because we see them every day, they have also had an enormous impact in other, less public, fields. As a whole book could be written on the uses of computer graphics, for the sake of brevity this chapter concentrates upon 'hard' applications of graphics, such as scientific visualisation and image processing.

Computer-Aided Design (CAD)
Engineering has been transformed by computer graphics in the form of CAD. In the past drawings had to be painstakingly prepared by draughtspeople and mock-ups and prototypes of products built at every stage of development. With CAD, however, the product can be designed and tested within the computer, drastically reducing the time taken to get products from the draughting table to the market. CAD is used in all areas of engineering, from car design to electronics.
Not surprisingly, every CAD drawing is a vector graphic composed of objects that the designer can manipulate at will. Both individual objects and collections of objects can be tested: the designer can apply known inputs to an electronic circuit, for example, to see if it produces the required output, or a wing section could be put under stress in a 'virtual' wind tunnel by the computer. So long as the models of behaviour of objects like wing struts and electronic components are accurate then the testing precludes the production of physical mock-ups. When such physical objects are required, the computer can drive a machine to create a model for testing, a process which is both quick and totally accurate.
Because of the intense computational requirements of CAD it is usually carried out using high-end, dedicated workstations using fast processors and high-resolution monitors, although as micros continue to improve in performance it's likely that more CAD software will appear in the micro market.

Visualisation
Visualisation is the process of rendering data as images in order to enhance our understanding. The idea is not new: maps are visualisations of terrain, and since the early days of printing graphs and charts have been used to represent both empirical data and mathematical functions. Although visualisation is used in nearly all fields of human endeavour where computers may be used to process and display data, it's fair to say that the technique is used to the greatest extent in the commercial and scientific worlds.

Business Graphics
This term covers charts and graphs that represent economic statistics and which are generally used by managers in both in the public and private sectors. In particular, business visualisation is used in analysis, to extract meaning from data, and for presentation, to explain data to an audience. The use of images in economics considerably predates the computer age1, but of course computers enable the production of many kinds of highly effective and colourful charts from raw data in seconds; as a result a thriving market in presentation software exists (see the section on Presentation Packages in the chapter on Graphics Packages).



Figure 1
Common types of chart

Scientific Visualisation
In the world of science such are the vast volumes of data generated by modern technologies (such as meteorological satellites) that visualisation is not just a means of presenting the data but is very often essential in order to extract meaning and information2. Visualisation is a means of scientific enquiry and research, an analytical tool which is an integral part of the scientific process. It has had a dramatic impact on all natural sciences (and more than a few social sciences), and in some cases has actually created new fields of investigation, for example Fractal Geometry.
Visualisation takes many forms, from simple graphs to 3-D models such as terrain maps, [DIAGRAMS/EXAMPLES] the main purpose being to allow scientists to use the powerful information processing capabilities of their eyes to pick out salient information from the images presented to them. Whilst sometimes the image is analogous to the data being modelled visualisation also allows the 'unviewable' to be viewed [DIAGRAM: 3-d math function]such as complex mathematical functions. Viewing the function as a scene of 'mountains' and 'valleys' can enable the mathematician to discern properties of the function which would not be apparent using traditional data analysis.

Geographical Information Systems (GIS)
A GIS is a form of visualisation based on real-world empirical data about human and physical geography. The database (usually relational) is central to a GIS, together with cartographic (map-making) and of course graphics functionality. Data can be gathered from primary sources, such as aerial or satellite photographs, or extracted from secondary sources, such as census returns and historical archives. A major characteristic of GIS is the overlaying of topographic data - representing the physical environment - by thematic data - representing characteristics of that environment. For example, overlaying a map of the Lake District with geological data from the last Ice Age can produce a graphic of the icesheet distribution over the area. [DIAGRAM]

Image Processing
Put simply, Image Processing (IP) is the processing of one raster image (bitmap) by a computer to produce another. As raster images are composed of pixels, each of which has a digital value, they are extremely amenable to manipulation simply through arithmetic operations on the pixel values. As a simple example, multiplying each pixel in a greyscale image by 2 would double the intensity3 of that image.The purpose of IP is to enhance a captured real-world image to extract information of interest to the viewer. Just a few of the many image enhancements that can be applied are:

Smoothing. This tends to average out pixel values, has the effect of 'turning down the contrast' and removing image 'noise'.
Edge Detection. Sharp changes in pixel values in a small area can be emphasised so as to bring out previously unnoticed boundaries in the image.
Correlation. Different areas in the image are compared to find similarities, which may indicate the presence of similar objects.
False Colour. Many captured images, and particularly those from satellites, are in grey scale. Differences in grey level are harder for us to perceive than colour changes, so arbitrarily allotting a different colour to each level can make the image easier to understand.
Pattern Recognition. An image, or more usually parts of it, can be compared to stored images to find similarities. This is the basis of OCR (Optical Character Recognition) and Computer Vision.

[DIAGRAM: satellite image, and same image with IP applied]
IP finds uses in a wide range of fields, including:

· archaeology
· art
· astronomy
· biology
· computer vision
· geology
· medicine
· meteorology
· photography
· warfare

to name but a few. Any area that uses real-world captured images will have a use for IP.

Fractals

Fractal images have had a strong aesthetic impact in the world outside computing and science, particularly in youth culture. These eerily beautiful computer graphics, with their bright colours and complex whorls which seem to pull the eye deeper and deeper into the innate complexity of the image, are purely the product of computer power applied to simple, recursive mathematical equations. More importantly Fractal Geometry has reconnected pure mathematics to the natural sciences by allowing a mathematical treatment of natural processes, such as the formation of clouds, the growth of a plant, or the fluctuations of animal populations.
The term 'fractal', coined by the mathematician Benoit Mandelbrot, comes from 'fractional dimension', a curious concept when we are used to dealing in one, two and three dimensions. Most objects in the natural world possess fractional dimensions: coastlines, landscapes and clouds have fractional dimensions of around 1.2, 2.2 and 3.3 respectively.4
Fractals possess the properties of self-similarity and infinite complexity within a finite area. This can be seen by asking the question, 'How long is the coastline of Britain?', the answer to which is that it depends on the unit of measurement used. The more the coastline is magnified the more detail emerges, detail which exhibits similarities to larger-scale features. As the measuring unit becomes smaller so the length increases, so that in effect the coastline of Britain has infinite length within a finite area5.
In contrast to traditional Euclidean geometry, where a shape or curve is defined by an equation (eg y = 2x + 3), a fractal curve is defined by an algorithm, a procedure. For example, the famous Mandelbrot Set (Plates *** to ***) is defined by the formula z ®z2 + c where z is a complex number (see Appendix A) and c a constant. The expression z2 + c is computed, and the resulting value of z is used for the next computation of the expression, and so on until z reaches a particular value or a set number of iterations has occurred; the final value of z determines the position and colour of the plotted point. Thus each point is the result of many tens of iterations of z2 + c, so that on a 640 x 480 VGA monitor with 307,200 pixels the number of calculations runs into the millions.
It's because of these vast numbers of calculations required to produce one image that Fractal Geometry is such a young field. Although some mathematicians, even in the 19th century, had suspected that non-linear equations behaved strangely, it was only the advent of fast computers capable of handling the millions of necessary computations and displaying the results graphically which enabled fractals to be seen at last. Now Fractal Geometry and its close relative Chaos Theory - the study of non-linear feedback systems that are highly sensitive to initial conditions - are used as analytical tools in many natural and social sciences, from metereology to economics.

Education & Training
If a picture is worth a thousand words, there can be few more obvious uses for it than in the area of imparting knowledge to others. As noted previously (see the chapter on Multimedia) text is not a very natural teaching medium for us. Not only does it have a low information density but it also results in processing overheads for the brain, insofar as it is a symbolic representation of information that has to be decoded. Still and moving images, on the other hand, usually require no decoding (although they need a lot of interpretation) and have high information densities. Moreover, the human visual system has evolved to be highly efficient at information gathering and processing, so the presentation of information as graphics takes advantage of this natural ability. (This is by no means to deride text - plainly it would not be possible to put across the information in this book purely graphically - but simply to recognise its limitations.)
It would not be accurate to say as yet that computer graphics has qualitatively changed the delivery of education and training, although it has certainly significantly enhanced the quality of teaching. In contrast to science, where new technologies open up new avenues of research and change actual scientific practice, the practice of teaching changes slowly and is driven by theories of learning as well as more contingent socio-economic factors. Primary of these factors, of course, is the availability of funding, and this has meant that the use of computer graphics in public education has lagged behind private training. Corporations are willing to spend large sums on high-end training facilities because of the real productivity gains that can be realised and measured in cash terms. This contrasts with public education where 'productivity' is a more slippery concept and where the information being conveyed is of a different nature from training: broadly speaking education is about techniques and concepts - the How and Why - whereas training concentrates narrowly on the How.
This concentration upon technique makes training much more amenable to the use of computer graphics than education. It's easy to use graphics to train someone to put together an electronic circuit, but rather harder to explain the quantum mechanics that enable the semiconductors in the circuit to work. Multimedia in particular has great potential for training, and is already being extensively used in the form of laserdisks (see the chapter on Multimedia).
However, there are areas of education where graphics and multimedia can, and are, being put to good use. In the sciences visualisation techniques, as discussed earlier in this chapter, can as easily be used for teaching as analysis, and in many fields images are essential to grasp fundamental concepts - the structure of a molecule in chemistry, or the process of cell division in biology, for example. The teaching of history, particularly of the 20th century, could benefit from the use of the vast amount of still and moving images in the archives, as could the teaching of Art.
Hypermedia systems can be particularly useful in higher education. As noted in the chapter on Multimedia, one of the characteristics of hypertext and hypermedia is that they put the user back in control. In an educational setting this translates to student-centred learning, where the students use courseware - educational software - to learn at their own pace in their own time. This fits in with modern views of learning, which place emphasis on the engagement of the student in the learning process, rather than as a passive receptacle for knowledge. Moreover, the potential of multimedia for creating simulations based on real-world situations (that have been captured on video, for instance) increases the interaction of the student with the learning material and encourages active thinking and problem-solving6.
A logical extension of student-centred learning is distance learning, and as recent rapid advances in telecommunications have enabled the real-time transmission of audio and video on ordinary telephone lines the science-fiction scenario of a class of students spread over hundreds or thousands of miles being addressed by one lecturer, or accessing multimedia courseware, will soon become reality. (See the section on Videoconferencing in Chapter ***.)
There are a few subjects which would not benefit substantially from the use of computer graphics in their teaching: it's difficult to see how courses in such areas as English Literature, Philosophy, Music, or Religion, could be much enhanced by images whether moving or still. These, though, are the exceptions rather than the rule, and in most subjects the judicious use of graphics can only improve the teaching material. If nothing else, the inclusion of interesting images and sounds can make learning fun, which of course has a positive effect on the student's attitude and thus on knowledge uptake.