.sp 10
.ce 1
Computer Vision:
.ce 1
A Practical Introduction
.bp
.sp 4
.ce 1
Computer Vision:
.ce 1
A Practical Introduction
.sp 3
.ce 1
J. R. Parker
.ce 1
The University of Calgary
.ce 1
Department of Computer Science
.ce 1
Calgary, Alberta, Canada
.sp 18
.nf
	New York	Chichester	Brisbane	Toronto		Singapore
.bp
.sp 21

Copyright   (2016) by J Parker.
.sp 1
All rights reserved. Published simultaneously in Canada
.sp 1
Library of Congress Cataloging in Publication Data:
.sp 1
Parker, James Robert.
.in 4
Computer Vision: A Practical Introduction
.sp 1
Bibliography: p. 325
.br
Includes index.
.br
1. Computer Vision.  2. Image Processing  I. Title
.in 0
TA 1632.P92   1992  629.8'92  92-28763
.sp 3
Printed in the United States of America
.sp 2
10 9 8 7 6 5 4 3 2 1
.bp
.fi
Preface
.sp 5
.ls 2
.sp 1
.in 4
.ti 4
Computer vision has become a high profile subject in recent times.
The reason is clear enough - for those of us who are sighted, most of
the information we acquire day to day is through vision. We are
highly picture oriented beings, as demonstrated by the prevalance of
television, motion pictures, magazenes, and books. Even in science
much of the data has a visual aspect: photographs, spectrograms,
chromatography images, and a host of other visual means of accessing data
pervade all scientific disciplines. Indeed, when data is not visual
we tend to make it so by drawing graphs and charts.
.sp 1
.ti 4
Almost everyone has seen a computer enhanced picture in the form of
a Voyager or Pioneer spacecraft image of one of the outer planets.
Others have been exposed to computer images through the recent use
of the 'morphing' technique used in some recent music videos.
Both have increased the public awareness of the use of computers
in manipulating pictures, although neither example actually represents
computer vision. Voyager images more properly illustrate
image processing,
in which mathematical and other techniques are used to improve the quality
of an image for some particular purpose. For example, the Voyager images
were actually black and white, and were subject to many sources of noise
and distortion. Image processing methods applied to a sequence of such
images allowed the beautifully clear full color pictures that appeared
on television and in print. In a similar vein, the digitally morphed
sequences from the Michael Jackson video were generated by applying a
meld of image processing and computer graphics to films of carefully
choreographed motions by actors. The result is a smooth transition between
frames on the film, where it is impossible to tell which parts of the pictures
are real and which are computer generated.
.sp 1
.ti 4
Computer vision is involved with the extraction of information from images, and
in the identification and classification of objects in an image. A friend of
mine has referred to it as 'un-graphics', and this is very near the mark.
Computer vision systems are used to recognize faces and signatures, match
fingerprints, inspect parts on assembly lines, and guide robots in assembly tasks.
Of course there is an aspect of image processing in all computer vision systems.
Images need to be enhanced to emphasize important features before the vision
aspect can be performed; and there is an aspect of computer graphics in all
vision systems, since the images must be displayed, windowed, and scaled.
Because of this, any book on computer vision must contain some elements of
these other fields.
.sp 1
.ti 4
This book was written for the beginner in computer vision. It assumes only
basic mathematics such as would be taught in high school and early college
math courses. It does assume a knowledge of computers and programming
terminology, and examples are coded in the C language so a facility in C
would be useful. The intent is to provide a useful, if not complete, and
practical, if not completely rigorous, knowledge of basic computer vision.
While it could be used as a text in some computer science courses, it
is intended to be used by the general computing public, and by students
of subjects in which computer vision is a useful tool. Examples are drawn
from astronomy, geography, geology, biology, and medicine.
.sp 1
.ti 4
This book also contains code for a low-level vision system which I call Alpha
(optimistically thinking that there may one day be a Beta), which was developed
on a Sun workstation, but which is intended to compile and run on an IBM
PC running Dos and Borland C. The diskette that accompanies this book has the
source code for Alpha as well as a set of images which should be useful for
conducting experiments and for demonstrating the Alpha software.
The sample images lend to the practical aspect of this book. The reader
has immediate access to images with which to test their own ideas and
software.
.sp 1
.ti 4
The organization of this book is, I think, unique. Rather than being organized
by type of operation (enhancements, restorations, etc), this book is organized
by complexity. Following the introduction we start with an examination of the
simplest kind of image, one with only black or white. This chapter covers many of
the useful operations on and measurements of bi-level pictures. Next we look
at images having many levels of grey, and discuss enhancements and measurements,
including ways to convert to bi-level form. Next is a chapter on high level
representation of image objects, and the matching of patterns. Finally there
are a collection of chapters each dealing with a different real problem
in low level computer vision. After the problem is explained an attempt will
be made to solve it using existing Alpha code and such new approaches as
are needed for the particular case.
I feel that this organization allows a clear flow of ideas, and the final
chapters lend a practical aspect to the book not found elsewhere.
.sp 1
.ti 4
Of course there is a lot of material not covered by this book. Restoration
is not covered at all, and filters, a vast subject, are only covered superficially.
Three dimensional vision is avoided, and so is the subject of color, since
these areas deserve more coverage than is possible in a book at this level.
The basic idea here is to present to scientists, engineers, and interested
computer users of all stripes the essential ideas surrounding computer vision
in an immediately usable and understandable fashion. The mathematics which
is so key to the development of the discipline is avoided in favour of
a more intuitive approach, and the hope is that sufficient interest will
be generated by reading this book that the reader will then pursue the subject
further, in more depth.
.sp 1
.ti 4
The text and many of the computer generated figures were prepared using a
Sun 3/60 and Sparcstation 2, and code for Alpha was originally designed
using these machines. The software for the diskettes was prepared on an
Packard-Bell notebook computer with a VGA card.
.sp 1
.ti 4
Significant contributions to this effort were made by my wife, Katrin, who
suffered through the hours and the mood swings, and my children Adam and Bailey,
who did the same and posed for some sample images. 
Thanks to: Mike Williams for
editorial help and the use of his IBM computer;
John Heerema for the use of IBM PC hardware;
and to
the authors of the XV software for the Sun, without which I would have
been lost.
Great thanks for wonderful sample images
goes to:
Gaston Groisman, former graduate student and current friend;
Cullen Jennings, current graduate student and ISEF expert;
Dr. Ata Sarajedini of Yale University;
Dr. Roland Auer and Dr. Fred Biddle of the University of Calgary
medical school Pathology Department;
Dr. R. MacLaughlan from the Biology Department at the University of Calgary;
Dr. N. Wardlaw from the Geology Department;
Dr. M. Hawes and Dr. J. Engsberg of the University of Calgary 
Human Performance Laboratory;
Dr. R. Callaghan and Dr. J. Helmer of the
University of Calgary Archaeology Department;
the EECE Department of the University of New Mexico for the 
Khoros images and software;
Dr. S. Kwok of the Department of Physics and Astronomy for the Cygnus
X-ray image; Big Hill Veterinary Services in Cochrane, Alberta for
the animal X-rays;
and to NASA  for the Voyager and Magellan data.
.sp 4
Jim Parker
.br
Calgary, Alberta, Canada
.br
November, 1992.
.bp
.fi
.ls 1
Contents
.sp 5

Preface
.sp 1
.nf
1. Introduction
.in 3
1.1 Images as digital objects
1.2 Image storage and display
1.3 Image acquisition
1.4 Image types and applications
Summary
Exercises
.in 0
.sp 1
2. Bi-Level Images
.in 3
2.1 Usefulness of bi-level images
2.2 Connectivity and geometry
2.3 Measurable properties
2.3.1 Area
2.3.2 Perimeter
2.3.3 Length
2.3.4 Moments - center of mass
2.3.5 Simple shape
2.3.6 Derivative and Complex Shape Measures
2.4 Operations on bi-level images
2.4.1 Boundary enhancement
2.4.2 Erosion and dilation
2.4.3 Skeletonization
2.4.4 Chain codes
2.4.5 Run Length Encoding
Summary
Exercises
.in 0
.sp 1
3. Grey Level Images
.in 3
3.1 Thresholding
3.2.1 Selecting A Single Threshold
3.2.2 Selecting Multiple Thresholds
3.2 Grey level modification
3.3 Lines and edges
3.4 Geometric operations
3.5 Noise
3.6 Color
Summary
Exercises
.in 0
.sp 1
4. Representing and Recognizing Objects
.in 3
4.1 Features
4.2 Statistical Pattern Analysis
4.3 Decision Functions
4.4 Template Matching
4.5 Structural Pattern Recognition
4.4.1 Representing Relationships
4.4.2 Identifying Components
Summary
Exercises
.in 0
.sp 1
5. Counting and Classifying Objects
.in 3
5.1 Counting simple cells
5.2 Classifying seeds
5.3 Classifying galaxies
5.4 Defects in printed circuit boards
Summary
Exercises
.in 0
.sp 1
6. Computer Readable Codes
.in 3
6.1 The bar code
6.2 Fonts for machine readable text
6.3 The general OCR problem
Summary
Exercises
.in 0
.sp 1
7. Scientific Images
.in 3
7.1 Chromatography
7.2 Star images
7.3 Synthetic Color Voyager Images
7.4 Computer Vision In Archaeology
Summary
Exercises

.in 0
.sp 1
Appendix A: The Alpha Vision System
.sp 1
Appendix B: Systems
.sp 1
Appendix C: Bibliography
.sp 1
Index
.bp
.sp 15
.ce 2
"We shall not all sleep; but 
we shall all be changed."
.ce 1
II Cor. 16:19
