• No results found

1.6.1 Journals and magazines

As in any academic subject, there are many sources of literature. The professional magazines include those that are more systems oriented, like Image Processing and Advanced Imaging. These provide more general articles, and are often a good source of information about new computer vision products. For example, Image Processing often surveys available equipment, such as cameras and monitors, and provides a tabulated listing of those available, including some of the factors by which you might choose to purchase them. Advanced Imaging is another professional journal that can cover material of commercial and academic interest. There is a wide selection of research journals – probably more than you can find in your nearest library unless it is particularly well stocked. These journals have different merits: some are targeted at short papers only, whereas some have short and long papers; some are more dedicated to the development of new theory whereas others are more pragmatic and

function inverted=invert(image)

%Subtract image point brightness from maximum %

% Usage: [new image]=invert(image) %

% Parameters: image-array of points %

% Author: Mark S. Nixon %get dimensions

[rows,cols]=size(image);

%find the maximum maxi=max(max(image));

%subtract image points from maximum for x=1:cols %address all columns

for y=1:rows %address all rows inverted(y,x)=maxi-image(y,x); end

end

focus more on practical, working, image processing systems. But it is rather naive to classify journals in this way, since all journals welcome good research, with new ideas, which has been demonstrated to satisfy promising objectives.

The main research journals include: IEEE Transactions on: Pattern Analysis and Machine

Intelligence (in later references this will be abbreviated to IEEE Trans. on PAMI); Image Processing (IP); Systems, Man and Cybernetics (SMC); and Medical Imaging (there are

many more IEEE transactions, some of which sometimes publish papers of interest in image processing and computer vision). The IEEE Transactions are usually found in (university) libraries since they are available at comparatively low cost. Computer Vision

and Image Understanding and Graphical Models and Image Processing arose from the

splitting of one of the subject’s earlier journals, Computer Vision, Graphics and Image

Processing (CVGIP), into two parts. Do not confuse Pattern Recognition (Pattern Recog.)

with Pattern Recognition Letters (Pattern Recog. Lett.), published under the aegis of the Pattern Recognition Society and the International Association of Pattern Recognition, respectively, since the latter contains shorter papers only. The International Journal of

Computer Vision is a more recent journal whereas Image and Vision Computing was established

in the early 1980s. Finally, do not miss out on the IEE Proceedings – Vision, Image and

Signal Processing and IEE Proceedings – Digital Techniques.

Some of the journals are now on-line but usually to subscribers only, in the UK through Ingenta through BIDS (you need an account at Bath Information and Data Services at

http://www.bids.ac.uk/). Academic Press appear to be mostly on-line now, including

Computer Vision and Image Understanding, Graphical Models and Image Processing and Real-Time Imaging at http://www.apnet.com/www/journal/iv.htm, http:/ / w w w . a p n e t . c o m / w w w / j o u r n a l / i p . h t m, and h t t p : / / w w w .

academicpress.com/rti respectively.

1.6.2 Textbooks

There are many textbooks in this area. Increasingly, there are web versions, or web support, as summarised in Table 1.4. This text aims to start at the foundation of computer vision, and ends very close to a research level. Its content specifically addresses techniques for image analysis, considering shape analysis in particular. Mathcad and Matlab are used as a vehicle to demonstrate implementation, which is rarely considered in other texts. But there are other texts, and these can help you to develop your interest in other areas of computer vision.

This section includes only a selection of some of the texts. There are more than these, some of which will be referred to in later chapters; each offers a particular view or insight into computer vision and image processing. The introductory texts include: Fairhurst, M. C.: Computer Vision for Robotic Systems (Fairhurst, 1988); Low, A.: Introductory

Computer Vision and Image Processing (Low, 1991); Teuber, J.: Digital Image Processing

(Teuber, 1993); and Baxes, G. A.: Digital Image Processing, Principles and Applications (Baxes, (1994) which includes software and good coverage of image processing hardware. Some of the main textbooks include: Marr, D.: Vision (Marr, 1982) which concerns vision and visual perception (as previously mentioned); Jain, A. K.: Fundamentals of

Computer Vision (Jain, 1989) which is stacked with theory and technique, but omits

implementation and some image analysis; Sonka, M., Hllavac, V. and Boyle, R. Image

Processing, Analysis and Computer Vision (Sonka, 1998) offers more modern coverage of

implementation but omitting some image processing theory; Jain, R. C., Kasturi, R. and Schunk, B. G.: Machine Vision (Jain, 1995) offers concise and modern coverage of 3D and motion (there is an on-line website at http://vision.cse.psu.edu/ with code and images, together with corrections); Gonzalez, R. C. and Wintz, P.: Digital Image

Processing (Gonzalez, 1987) has more tutorial element than many of the basically theoretical

texts; Rosenfeld, A. and Kak, A. C.: Digital Picture Processing (Rosenfeld and Kak, 1982) is rather dated now, but is a well-proven text for much of the basic material; and Pratt, W. K.: Digital Image Processing (Pratt, 1992) which was originally one of the earliest books on image processing and, like Rosenfeld and Kak, is a well-proven text for much of the basic material, particularly image transforms. Despite its name, the recent text called

Active Contours (Blake, 1998) concentrates rather more on models of motion and deformation

and probabalistic treatment of shape and motion, than on the active contours which we shall find here. As such it is a more research text, reviewing many of the advanced techniques to describe shapes and their motion. A recent text in this field, Image Processing – The

Fundamentals (Petrou, 1999) surveys the subject (as its title implies) from an image

processing viewpoint covering not only image transforms, but also restoration and enhancement before edge detection. The latter of these is most appropriate for one of the major contributors to that subject. Also, Kasturi, R. and Jain, R. C. (eds): Computer Vision:

Principles (Kasturi, 1991a) and Computer Vision: Advances and Applications (Kasturi,

1991b) presents a collection of seminal papers in computer vision, many of which are cited in their original form (rather than in this volume) in later chapters. There are other interesting edited collections (Chellappa, 1992), one edition (Bowyer, 1996) honours Azriel Rosenfeld’s many contributions.

Books which include a software implementation include: Lindley, C. A.: Practical

Image Processing in C (Lindley, 1991) and Pitas, I.: Digital Image Processing Algorithms

(Pitas, 1993) which both cover basic image processing and computer vision algorithms. Table 1.4 Web textbooks and homepages

This book’s Southampton U http://www.ecs.soton.ac.uk/~msn/book/

homepage

CVOnline Edinburgh U http://www.dai.ed.ac.uk/CVonline/

Ad Oculos Imaging Source http://www.theimagingsource. com/prod/link/adoculos.htm

Image Processing Delft University http://www.ph.tn.tudelft.nl/

Fundamentals Courses/FIP/noframes/fip.html

World of Wolfram Research http://mathworld.wolfram.com

Mathematics

Numerical Recipes Cambridge http://www.nr.com/

University Press

The Joy of Visual York U http://www.yorku.ca/research/

Perception vision/eye/thejoy.htm

Machine Vision Penn State http://vision.cse.psu.edu/

homepage

Active Contours Oxford U http://www.robots.ox.ac.

Parker, J. R.: Practical Computer Vision Using C (Parker, 1994) offers an excellent description and implementation of low-level image processing tasks within a well-developed framework, but again does not extend to some of the more recent and higher level processes in computer vision and includes little theory though there is more in his later text Image

Processing and Computer Vision (Parker, 1996). A recent text Computer Vision and Image Processing (Umbaugh, 1998) takes an applications-oriented approach to computer vision

and image processing, offering a variety of techniques in an engineering format, together with a working package with a GUI. One recent text concentrates on Java only, Image

Processing in Java (Lyon, 1999), and concentrates more on image processing systems

implementation than on feature extraction (giving basic methods only). As already mentioned, the newest textbook (Efford, 2000) offers Java implementation, though it omits much of the mathematical detail making it a lighter (more enjoyable?) read. Masters, T.: Signal and

Image Processing with Neural Networks – A C++ Sourcebook (Masters, 1994) offers good

guidance in combining image processing technique with neural networks and gives code for basic image processing technique, such as frequency domain transformation.

There are now a number of books on the web as given in Table 1.4. This book’s homepage has a link to these web-based texts, and will be kept as up to date as possible. The CVOnline site describes a great deal of technique, whereas the Ad Oculos page describes the book that supports the software. Image Processing Fundamentals is a textbook for image processing. The World of Mathematics comes from Wolfram research (the distributors of Mathematica) and gives an excellent web-based reference for mathematics. Numerical

Recipes is one of the best established texts in signal processing. It is beautifully written,

with examples and implementation and is on the web too. The Joy of Perception gives you web access to the processes involved in human vision (and the worst title?).

Other textbooks include: Russ, J. C.: The Image Processing Handbook (Russ, 1995)

which contains much basic technique with excellent visual support, but without any supporting theory, and has many practical details concerning image processing systems; Davies, E. R.:

Machine Vision: Theory, Algorithms and Practicalities (Davies, 1994) which is targeted

primarily at (industrial) machine vision systems but covers much basic technique, with pseudocode to describe their implementation; and Awcock, G. J. and Thomas, R.: Applied

Image Processing (Awcock, 1995) which again has much practical detail concerning image

processing systems and implementation.

1.6.3 The web

The web entries continue to proliferate. A list of web pages is given in Table 1.5 and these give you a starting point from which to build up your own list of favourite bookmarks. All these links, and more are available at this book’s homepage http://www.ecs.

soton.ac.uk/~msn/book/). This will be checked regularly and kept up to date.

The web entries in Table 1.5 start with the Carnegie Mellon homepage (called the Computer Vision Homepage). The Computer Vision Online CVOnline homepage has been brought to us by Bob Fisher from the University of Edinburgh. There’s a host of material there, including its description. Their group also proves the Hypermedia Image Processing Website and in their words: ‘HIPR2 is a free www-based set of tutorial materials for the 50 most commonly used image processing operators. It contains tutorial text, sample results and Java demonstrations of individual operators and collections.’ It covers a lot of basic material and shows you the results of various processing options. A big list of active groups can be

found at the Computer Vision homepage and searchers like Google or Altavista can be a boon when trawling the web. If your university has access to the web-based indexes of published papers, the ISI index gives you journal papers (and allows for citation search), but unfortunately including medicine and science (where you can get papers with 30+ authors). Alternatively, Compendex and INSPEC include papers more related to engineering, together with papers in conferences, and hence vision (INSPEC in particular), but without the ability to search citations. Citeseer is increasingly useful. Two newsgroups can be found at the addresses given in Table 1.5 to give you what is perhaps the most up-to-date information.