Page 25 -

Preface xxiii

clear a description as possible. The software that accompanies the descriptions
is certainly an alternative to the math, and gives a step-by-step description of
the algorithms.
I have deleted some material completely from the first edition. There is no
longer a chapter on wavelets, nor is there a chapter on genetic algorithms.
On the other hand, there is a new chapter on classifiers, which I think was
an obvious omission in the first edition. A key inclusion here is the chapter
on the use of parallel programming for solving image-processing problems,
including the use of graphics cards (GPUs) to accelerate calculations by factors
up to 200. There’s also a completely new chapter on content-based searches,
which is the use of image information to retrieve other images. It’s like saying,
‘‘Find me another image that looks like this.’’ Content-based search will be an
essential technology over the next two decades. It will enable the effective use
of modern large-capacity disk drives; and with the proliferation of inexpensive
high-resolution digital cameras, it makes sense that people will be searching
through large numbers of big images (huge numbers of pixels) more and more
often.
Most of the algorithms discussed in this edition can be found in source
code form on the accompanying web page. The chapter on thresholding alone
provides 17 programs, each implementing a different thresholding algorithm.
Thinning programs, edge detection, and morphology are all now available on
the Internet.
The chapter on image restoration is still one of the few sources of practical
information on that subject. The symbol recognition chapter has been updated;
however, as many methods are commercial, they cannot be described and
software can’t be provided due to patent and copyright concerns. Still, the
basics are there, and have been connected with the material on classifiers.
The chapter on parallel programming for vision is, I think, a unique feature
of this book. Again using downloadable tools, this chapter shows how to link
all the computers on your network into a large image-processing cluster. Of
couse, it also shows how to use all the CPUs on your multi-core and, most
importantly, gives an introductory and very practical look at how to program
the GPU to do image processing and vision tasks, rather than just graphics.
Finally, I have provided a chapter giving a selection of methods for use
in searching through images. These methods have code showing their imple-
mentation and, combined with other code in the book, will allow for many
hours of experimenting with your own ideas and algorithms for organizing
and searching image data sets.
Readers can download all the source code and sample images mentioned in
this book from the book’s web page — www.wiley.com/go/jrparker.You can
also link to my own page, through which I will add new code, new images,
and perhaps even new written material to supplement and update the printed
matter. Comments and mistakes (how likely is that?) can be communicated