“Foundations and Trends in Computer Graphics and Vision”
Brian Curless, Luc Van Gool, and Richard Szeliski.
Now Publishers, Inc.
Reviewed by: Larry O’Gorman
In the course of my job as editor of the IAPR Newsletter, I am always on the lookout for newly published books to review here. I saw a couple of book announcements a few months ago and requested review copies from the publisher. I was informed that these books came from individual issues in a journal, Foundations and Trends in Computer Graphics and Vision published by Now Publishers. The charter of this journal is to publish survey and tutorial articles. I found this very interesting and somewhat rare (ACM Computing Surveys has a similar charter). Since I’ve long encouraged survey, tutorial, and comparison papers as a necessary complement to “new material” papers, I decided to look into the journal more deeply and convey my thoughts here.
Foundations and Trends in Computer Graphics and Vision began publication in 2005 and publishes 4 issues per year. Each issue contains only one article. Issue page lengths are targeted at 100 pages, but one issue I saw reaches 178. Since the issues are book-length, they are available both by journal subscription and by purchase as individual books. As for many new publications, both journal and books are available in paper or online (html and pdf) formats. The editors-in-chief of this journal are Brian Curless, Luc Van Gool, and Richard Szeliski.
The issues published to date are:
“Monocular Model-Based 3D Tracking of Rigid Objects”, by V. Lepetit and P. Fua (Vol. 1-1, 2005)
“Computational Studies of Human Motion”, by David Forsyth, Okan Arikan, Leslie Ikemoto, James O’Brien, Deva Ramaman (Vol 1-2/3, 2005)
“Object Categorization”, by Axel Pinz (Vol. 1-4, 2005)
“Image Alignment and Stitching: A Tutorial”, by Richard Szeliski (Vol. 2-1, 2006)
“Mesh Parameterization Methods and Their Applications”, by Alla Sheffer, Emil Praun and Kenneth Rose (Vol. 2-2, 2006)
“Image-Based Rendering”, by Sing Bing Kang, Yin Li, Xin Tong, Heung-Yeung Shum (Vol. 2-3, 2006)
“A Stochastic Grammar of Images”, by Song-Chun Zhu and David Mumford (Vol. 2-4, 2006)
Upcoming issues include: “Deterministic methods for MRFs”, “Passive 3D Reconstruction”, “Computational Symmetry”, “Recognizing and Learning Object Categories”, “The Appearance of Human Skin: A Survey”, and “Local Feature Extraction”.
I examined a couple of issues in more depth and report on them here. The 1:4 (2005) issue was titled, “Object Categorization”, written by Axel Pinz. It is a 100-page tutorial article with 210 references. Since the article starts from fairly basic principles, it is appropriate for a student who has done a course in signal/image processing and/or pattern recognition, and is perhaps interested in learning background for doing an undergraduate project or Master’s thesis in this area.
Chapter 1 consists of a problem statement (assigning a specific object to a certain category by extraction and use of visual features), a short history of the field, and some potential applications. Chapter 2 describes categorization as related to a number of methods, including classification, learning, and datasets. Chapter 3 describes what it calls the “building blocks”, which include concepts that follow a logical processing sequence: pixels, edges, scale-space, moments, segments, recognition, and learning. Finally, Chapter 4 leaves behind general teaching to describe the author’s own prototype system for categorization. This includes region- and boundary-based image categorization. Although this article does not deal deeply with any particular aspect of object categorization, many students and practitioners will find this the right level of detail to understand some of the topic’s methods and challenges.
The other article I examined was in the 1:2/3 (2005) issue titled, “Computational Studies of Human Motion: Part 1, Tracking and Motion Synthesis”, by Forsyth, Arikan, Ikemoto, O’Brien, and Ramaman. While this is also a tutorial article, it has a different style than the previously described article in a number of ways. Quantitatively, it is much longer at 178 pages with 433 references. Its topic is narrower than above, but it treats it with more depth. In addition to describing the general field well, this article includes the authors’ own work throughout, as opposed to just the final chapter as above. This article is likely best aimed at graduate students with some background in computer vision who desire to learn more about the specific challenge of human motion. The authors describe this article as one half of a future book, where this half describes the earlier stages of motion tracking and synthesis, and the future half covers representation and motion generation.
Chapter 1 defines some fundamental notions of tracking including: by detection, using flow, and with probability. Chapter 2 discusses the relationships between 2- and 3-D. Chapter 3 digs deeper into human body tracking, discussing how to match the human body, generative and discriminative appearance models, human body part codebooks, and how to evaluate these approaches. Chapter 4 is on motion synthesis, how to take the features found from computer vision methods and begin synthesizing for the purpose of animation. I suspect this article and the subsequent book will be very useful for those whose work straddles the fields of computer vision and animation of the human body.
Back to the journal as a whole, there are a number of advantages that it brings to technical publishing that are, if not new, at least uncommon. Instead of forcing an article to fit into the roughly 20-page constraint of most technical journals, this journal allows substantially longer articles. This is very appropriate for surveys and tutorials. Articles are available both in series as a journal, which is most appropriate for libraries and teachers, and in book form, which is appropriate for practitioners only interested in a particular topic. Additionally, for authors, the copyright is retained by the authors.
Book Reviews Published in
the IAPR Newsletter
Applied Combinatorics on Words
by M. Lothaire
3D Shape Estimation and Image Restoration,
Exploiting Defocus and Motion Blur
by Paolo Favaro & Stefano Soatto
Human Identification Based on Gait
by Nixon, Tan and Chellappar
Mathematics of Digital Images
by Stuart Hogan
Advances in Image and Video Segmentation
Graph-Theoretic Techniques for Web Content Mining
by Schenker, Bunke, Last and Kandel
Handbook of Mathematical Models in Computer Vision
by Paragios, Chen, and Faugeras (Editors)
The Geometry of Information Retrieval
by van Rijsbergen
Biometric Inverse Problems
by Yanushkevich, Stoica, Shmerko and Popel
Correlation Pattern Recognition
by Kumar, Mahalanobis, and Juday
Pattern Recognition 3rd Edition
by Theodoridis and Koutroumbas
Dictionary of Computer Vision and
by R.B. Fisher, et. Al
Kernel Methods for Pattern Analysis
by Shawe-Taylor and Cristianini
Machine Vision Books
CVonline: an overview
The Guide to Biometrics by Bolle, et al
Pattern Recognition Books
Jul. ‘04 [pdf]