Clear and simple book, goes right to the guts. One minus is sloppy code. It's written by a scientist who is not a programmer. Implementations are slow and sometimes inaccurate. For example, function nays8 counts 0's in one thinning algorithm. Function snays count 0's in another thinnnig algorithm, while nays8 now count 1's!
Most of these comes from straightforward implementation of ideas. Sure enough, exactly this makes his code a good starting point in writing more effective programs.

It's a good book for people who is working in Computer Vision with C++ language. There are explanations of some algorithms and their respective implementation in C++, so it's very helpful. Furthermore, after each topic there is a bibliography list about this topic, which is better than have all bibiography in the last part of the book.
The only drawback is that the quality of the book is not good. The pages are very easy to tear.

This nifty book is unusual in several respects: it is written in an extremely clear and simple style; it presents good clear code, both in print and on the accompanying CD; and it presents a variety of different methods for achieving each given goal. For example, the book presents a half-dozen different algorithms (usually with code) for edge-detection, for thresholding, and for skeletonization. For the most part, the book does a good job of explaining the relative merits or suitabilities of the different algorithms.

It's important to say, though, that this book is not a comprehensive introduction to image processing issues and algorithms (such as Gonzales&Woods), nor is it a soup-to-nuts cookbook of code (such as Myler&Weeks' admittedly dated "Pocket Handbook"). Although the book is written in a style that students could understand, I wouldn't recommend it as a first book for novice students. It would be better suited to a non-specialist who nevertheless needs to accomplish something specific -- like thresholding or edge-detection -- in short order, and with an algorithm that's tuned to their particular application.

One last thing: this book will be VERY MUCH appreciated by anyone seeking a foundation in optical character recognition.

This book is a great source of code for the implementations of common and not so common algorithms used in image processing and computer vision. It is not meant to be a textbook on either subject. For that I suggest you turn to "Digital Image Processing" by Gonzalez and Woods for the subject of image processing and to "Computer Vision" by Linda Shapiro for computer vision.
The book starts out with the treatment of a common subject - edge detection - and provides code for various edge detectors including Canny, Shen-Casten, and Marr-Hildreth.
Chapter 2 does a pretty good job of explaining digital morphology, the various operations, and their uses.
Chapter 3 has a very brief introduction to gray-level segmentation, and then jumps into more advanced topics such as the method of iterative selection, entropy, fuzzy logic methods, and the method of Chow and Kaneko, among others. Most of the methods here are somewhat mathematically involved, and the author does a pretty good job of explaining the mathematics and showing resulting images that show the strengths of each method.
Chapter 4 is about the calculation and use of texture in image processing and computer vision. Again, the mathematics here can get involved and equations and illustrations are included to help make the concepts clear. Even fractal geometry and its use in texture is very briefly mentioned.
Chapter 5 is one of my favorites, since it covers the topic of skeletonization, which is not commonly covered in most vision or image processing books. The most common method is mentioned first - that of the medial axis transform. This is followed by a discussion of iterative morphological methods and also the use of contours. Finally, the chapter is rounded out with an explanation of the Zhang-Suen/Stentiford/Holt combined method and accompanying code. This method usually produces the best combination of good visual results and computational efficiency.
In Chapter 6, on image restoration, the author moves to the frequency domain and discusses the Fourier Transform, the FFT, and their uses in image restoration. The author provides his own image restoration system, the Eagle Restoration System, complete with code.
Chapter 7, on wavelets, is a very brief and not too satisfying introduction to the topic. The author does provide some mathematics and some code for some basic transforms, but does not provide much in the way of motivation.
Chapters 8 and 9, on optical character recognition and symbol recognition respectively, are two very good chapters on this subject. It provides the simple introductory information on recognizing printed alphabetic characters. It also talks about skew detection in the case of slanted lines of text. Chapter 9 expands the concepts with a very good section on neural networks and character recognition concluding with a system that recognizes printed music.
Chapter 10, the final chapter of the book, was an OK introduction to genetic algorithms and evolutionary computing. However, like the chapter on wavelets, I found it a bit thin.
The appendix contains a listing of all of the C code included on the CD and each program's purpose. Going through the individual programs was, for me, almost as educational as reading the book, since the code is very well laid out and commented. I highly recommend this book to individuals looking for implementations and discussions of certain image processing and computer vision algorithms.

It has source C code available, which I found is very helpful. It would be better if its code is written in C++.