PREFACE


The sixteen years since the publication of my first book on this topic have seen major expansion in the utilization of digital image processing. Algorithms that could run only on mainframe computers in the '60's and minicomputers in the '70's migrated to the desktop in the '80's. Personal computers transformed from something a few dedicated hobbyists built in the mid-70's into a common home office component. The jargon of personal computers became a universal language that bridged the oceans between the USA, Europe and Asia.

Public awareness of digital image processing has been greatly increased by video games, digital video special effects used in the entertainment industry and articles in the popular press. Present trends indicate a continuation of the explosive growth of digital image processing applications well into the next century.

Perhaps the most significant impact of digital image processing in the nineties will be in the area of applications to real-world problems. This book is aimed at the reader who intends to use this technology for research or commercial purposes. It also provides a foundation for those who seek to advance the state of the art.

While the scope and scale of digital image processing applications have changed dramatically, other aspects of the field have not. For example, many of the basic techniques that today perform reliably in practice are those that were first applied in the early days of digital imaging. While several exciting new theoretical areas have opened up, generally they build upon, rather than replace, what has served well in the past.

With the recent advances of computer technology, some of the issues treated in the earlier work are no longer of major concern. These are deemphasized in this book, while several relevant new topics have been included. New examples serve to further illustrate how the theory can be applied to the type of problems that commonly occur in industry and research.

Perhaps most significantly, a set of exercises and suggestions for projects complete each chapter. These have been selected to build the insight and understanding that is most useful to one endeavoring to apply this technology to problems of the real world. The majority of these emulate actual situations a professional faces working in this field. These are intended to give the reader a head start in gaining the insight that supplements a theoretical knowledge and can only come from the experience of solving real problems. In this author's estimation, one who not only knows how to solve these problems and conduct these projects, but has actually done in most of them, will be ready to take his or her place on the most productive image processing applications team.

For about twenty-five years, I have had the opportunity to observe the efforts of many individuals applying digital image processing techniques to problems offered by the real world. A few of these individuals have established an enduring track record of solid success on almost every attempt. They have consistently contributed innovative and effective solutions that creatively employ the tools of this discipline.

These highly productive individuals demonstrably hold several characteristics in common. One can venture to assume that these characteristics constitute a formula for success, to whatever extent such a thing can exist in this field.

Uniformly they have (1) a genuine interest in - even a fascination with - the technology involved, (2) a thorough understanding of the fundamentals of this highly multidisciplinary technology, (3) a conceptual type of understanding (as opposed to rote memorization or totally abstract theory), and (4) a knack for seeing problems visually, graphically and from more than one viewpoint. In line with this last point, they often find themselves hard pressed to explain their ideas without the aid of a graph or drawing.

This book is designed to help the reader develop the last three of these traits, and perhaps enhance the first as well. The selection of materials for inclusion (and, equally important, for omission), the examples used, the references cited, the exercises and project suggestions are all directed toward this goal.

In this field, mathematical analysis forms the stable basis upon which one can make definite predictions regarding the performance of a digital imaging system. In this treatment, however, mathematics is employed more as a faithful servant than as a ruthless master. The emphasis is on developing a conceptual understanding, and the analysis is used to support this goal.

The organization of this book generally follows that of the earlier text, simply because that particular flow of development proved to serve its purpose well. The level of mathematical complexity increases gradually through the first two parts of the book. While many readers have the background in mathematics required to begin the discussion with sampling theory and the Fourier transform, others do not.

More importantly, though, many of the most important concepts can be presented without the aid of advanced math. Thus we are able to avoid an additional element of complexity in the interests of making the learning process less burdensome and more interesting for all readers. As a general rule, topics receive attention in relation to their importance, rather than their complexity.

The field of digital image processing has now become so rich with technology that it is impossible to cover all aspects of it in a single volume of reasonable size. Thus we concentrate upon those techniques that prove most useful in practice and leave most of the mathematical proofs to the references. Constraints of paper and ink further make it impossible to include nearly as many example images as would be desirable. See [1] for an excellent source of these.

Part 1 presents several important concepts that do not require detailed mathematical analysis for a basic understanding. Part 2 addresses techniques that rely more heavily upon their mathematical underpinning, and it elaborates analytically upon certain concepts introduced in Part 1. Part 3 addresses applications more specifically than in earlier chapters.

A Note to Instructors. The development of this text has been accompanied by an accumulation of example digital images and problem solutions worked out in MathCAD™ [2] and WiT™ [3]. These are available from the World Wide Web site that supports this book (http://www.adires.com/~castlman/). The author can be reached on the Internet (castleman@adires.com or ken@castlman.org) or on Usenet (sci.image.processing).

A Note to Students. Digital imaging is a merger of several disciplines, and its nomenclature comes from many diverse fields. Often ordinary words are pressed into special new usage without warning. This can be quite confusing when it catches the reader unaware. Many of these specialized words are defined in Appendix 1. If the concept presented in a paragraph is not clear, check for a word that doesn't seem to fit. It there is one, look in the glossary or a dictionary for clarification. Frequent reference to the glossary and a dictionary is good insurance against difficulties in understanding the subject.

Image processing is best learned by a combination of study and application. One develops considerable insight by using the theory, working with actual imaging problems and image processing equipment. A balance between theory and practice keeps the subject interesting. Problems and projects are included at the end of each chapter for this purpose.


REFERENCES

1. G. A. Baxes, Digital Image Processing: Principles and Applications, John Wiley and Sons, New York, 1994. http://www.imageware.com/web/
2. MathSoft, Inc. 201 Broadway, Cambridge, MA 02139. http://www.mathsoft.com/
3. Logical Vision, Ltd., BC, Canada. http://www.logicalvision.com/


[Contents]

[Pixel Page] [Chap. 1->]


This page is maintained by Ken Castleman.
Copyright© 1995, Kenneth R. Castleman. All rights reserved.
Last modified 15 May, 2002


castleman@persci.com