The term non-photorealistic computer graphics has come to denote the area of scientific and technological endeavor dealing with the computer generation of images and animations that, generally speaking, appear to be made in part "by hand." Such images often resemble those that, for example, architects, industrial artists, or scientific illustrators produce to communicate more or less specific information, often accompanied by text. They are characterized by their use of randomness, ambiguity, or arbitrariness rather than completeness and adherence to the portrayed objects’ properties.
Non-photorealistic computer graphics involves all phases of processing that computer graphics in general uses. By far the most work has been has been done in what is denoted in this book by non-photorealistic rendering (NPR). It has its roots in early papers that appeared in the 1980’s (in particular [Strassmann, 1986] or [Sasada, 1987]). Two very influential papers were published at Siggraph 1990 ([Saito, 1990] and [Haeberli, 1990]), but the techniques they presented were still treated in isolation. In 1994, the contours of this new area began to emerge with the papers published at Siggraph ( [Winkenbach, 1994] and [Salisbury, 1994]) and Eurographics ([Strothotte, 1994]). These papers effectively broke open the dam by demonstrating the generality of the underlying principles.
After these publications in 1994, international conferences began having sessions devoted to non-photorealistic computer graphics. The first international symposium devoted solely to this topic was organized in Annecy, France, in June, 2000. By the time of this writing, it is estimated that the literature on this topic encompasses some 300 papers.
The time has become ripe for a systematic assessment of the literature. Having grown "organically," the methods and techniques that have been developed have lacked a uniform terminology and notation. The area has thus far been unstructured, making it increasingly difficult to identify and assess new open problems. Indeed, sometimes papers have even "reinvented the wheel," albeit in a different context and application concern. Indeed, this lack of a systematic study has led to the fact that at the time of this writing there is no single, all-encompassing tool for non-photorealistic computer graphics, neither in the market nor in research labs.
The book's use is threefold. First, it is intended to accompany a course within a computer science curriculum for students at the senior undergraduate or beginning graduate level. Preliminary drafts of the book were used by the authors for teaching such a course at the University of Magdeburg on three occasions (fall 1999, summer 2000, and fall 2000). The course encompasses 4 hours of lectures per week for a semester of 14 weeks. The students all had had at least one undergraduate computer graphics course covering the basics of 2D and 3D computer graphics. The students were expected to be proficient in a programming language.
The same course was taught by the authors at Simon Fraser University (Vancouver, Canada) as a two-week crash course with four hours of lectures per evening in weeks two and three of the semester (fall 2000). The students were given a take-home midterm in week 7 of the semester, and asked to submit a final project in week 13. This format worked well and enabled the students to take several other regular courses at the same time.
Students should be presented the material of the book in the order in which it is written. A sprinkling of the exercises at the end of the chapters should be given as homework. If there is not enough time to cover the whole book, some of the chapters can be thinned out. For example, Sections 2.3, 2.4, 3.3, 4.2, 4.3, 5.3, 7.3, 8.4, and 9.4 can be left out of the classroom but assigned as further reading without harming the students' basic understanding of the topic.
Second, the book will be useful to practitioners in the field. It contains a wealth of examples, particularly in the form of images, which the authors hope will excite the reader and motivate the use of non-photorealistic computer graphics. The methods introduced are explained in enough detail that programs can be written directly without major conceptual effort.
Computer graphics professionals wishing to get into the topic of non-photorealistic computer graphics either can read the chapters in order or, to save time in a first pass at the topic, can read more selectively. They should read Chapter 1, one of Chapters 2 through 5, Chapter 6, Chapter 9, and Chapter 10, if necessary skipping the sections mentioned for students.
The third use of the book is for reference by researchers in the field. It unifies the literature and introduces terminology. Wherever possible, the terminology introduced in the original papers is used within the book. However, in some cases, particularly where different articles use varying terminology, the book decides on one wording. The bibliographic references at the end of the chapters give the necessary pointers to the important publications.
In the case of researchers in the field of non-photorealistic computer graphics, the chapters can be read in just about any order because methods that are built upon are referenced appropriately. A comprehensive subject and author index aids in selective reading.
Should a course on a leading-edge topic such as NPR be part of a graduate degree program in computer science?
This question is really asking what is expected of computer science graduates. Presumably, students can no longer be endowed with an equally high level of specific engineering knowledge in all sub disciplines of computer science. Instead, there is an increasing demand for distilling what is being taught to core skills. These lie at the heart of the approach a computer scientist is to take when solving a problem. Such skills should be studied in the context of one another using any one of a number of example areas. The idea is that if these skills are mastered within one area, the graduate will be equipped with the ability to transfer the approach to other areas that may arise at their future workplace.
The area of NPR is one that exemplifies this approach. It takes an area of scientific endeavor that is treated with the methods and tools of theoretical, practical, and applied computing. The treatment of the subject matter as it appears in this book is to be exemplary for how computer scientists decompose problems into parts, bring individual solutions together again, and embed them in systems that actually help users carry out their tasks at hand.
Furthermore, one can observe that there has been a shift in the emphasis during the late 1990s toward providing graduates with a more user-centered view of their work. Whereas many areas within computer science, even within computer graphics, can be studied without ever carrying out empirical work with users, this book treats NPR as a subject area that begins with questions pertaining to what users really want to get out of using its methods and tools.
In keeping with the trend to more user-centered computing, there has been a tendency in recent years for Departments of Computer Science to devise new degree programs to meet the demands of the media industry. One example among many is the undergraduate and graduate program in Computational Visualistics offered at the University of Magdeburg. Here the emphasis is on methods and tools for visual communication, both from an algorithmic (computer science) and a user-centered (humanities) point of view. A course in NPR is of particular importance in this context because it demonstrates one aspect of the flexibility of graphical communication that will lie at the heart of Web-based systems in the first decade of the new millennium.
The material presented in this book draws on research results and the thoughts of many scientists. Our thanks go to Kees van Overveld for contributing his many deep insights into the topic in the final chapter of the book. A number of colleagues spent time with us in Magdeburg and provided their insights into the topic, among them Lyn Bartram, John Buchanan, Sheelagh Carpendale, Dave Forsey, and Simon Schofield. Many of the first author's Ph.D. students produced results that turned out to be instrumental in the development of this book. Thanks in this regard to Oliver Deussen, Bert Freudenberg, Frank Godenschweger, Nick Halper, Jörg Hamel, Knut Hartmann, Stefan Hiller, Axel Hoppe, Tobias Isenberg, Maic Masuch, Bernhard Preim, Andreas Raab, and Michael Rüger.
We wish to thank those persons who provided the support to make this book happen: the administrative, technical, and secretarial staff at our institute who keep things up and running, even under adverse workloads: Heiko Dorwarth, Volkmar Hinz, Petra Janka, Thomas Rosenburg, Petra Specht, and Sylvia Zabel; the students at the University of Magdeburg and Simon Fraser University who studied the topic with previous versions of the manuscript; and all of our colleagues around the world who gave us the copyrights to their images.
Finally, our particular thanks goes to the superbly professional staff at Morgan Kaufmann who turned our loose leaf pages into a book we are proud of: Mona Bühler, Diane Cerra, and Edward Wade.