Browse Books

Reviewer: Will Wallace

Until fairly recently, digitization and digital archival work have focused on physically smaller works such as visual arts, textual works, and other two-dimensional (2D) representational genres. The state of the art in digitization has been moving into the three-dimensional (3D) realm, while also tackling larger works such as sculpture and architectural works. This book is a new compendium of previously published articles and new material from the research group at the Ikeuchi Laboratory, University of Tokyo. The 23 chapters and appendices cover over 500 pages; 14 chapters were previously a part of various IEEE conference proceedings and transactions from 2001 to 2005; nine chapters of new material complement and nicely extend the previously reported research. The team at the University of Tokyo has been working on digitally archiving a range of objects, from problematic smaller works that are transparent or highly reflective, such as jewelry, stemware, and ceramic arts, to some of the most historically important and culturally significant works, including the 13-meter Kamakura Buddha in Japan, and the massive and architecturally complex Banyan Temple in Cambodia. The title of the volume, however, is somewhat misleading; the focus is really on the visual capture of these works through computer vision and computer graphics techniques, and on the many complex problems to ensure accurate shape, color, and 3D representation. Topics that are usually considered under the rubric of digital archiving, such as data representation standards, preservation, storage, retrieval, and display, are not covered in this volume. Still, the story is a fascinating one, and the volume is laid out very well, progressing logically through the initial image acquisition, rendering, and signal correction, to more advanced topics, primarily problems associated with the 3D nature of the objects they are digitizing. As the editors state in the introduction, the end goal of the work is "to develop methods to record and preserve current states" of significant cultural objects found outdoors. The book is divided into four parts that logically build on the previous sections. Part 1 consists of three chapters that describe the various sensors and techniques the research team has designed for data acquisition and digitization. Part 2 is comprised of eight chapters around the broad area of range data analysis and correction. It is quite interesting to read about the issues encountered while digitizing objects of this size in the outdoors, including all of the correction and novel techniques the team devised to correct the factors that degrade accurate representation: natural lighting, correction of data resulting from multiple perspective scans, and iterative algorithmic techniques to fill in missing data in 3D representation. Part 3 consists of six chapters on color analysis, and Part 4 has five chapters that detail broader issues associated with digitizing cultural heritage. The chapter on automated classification of Banyan faces from the Cambodian temple is quite interesting, revealing the complexity and variety of the figures throughout the temple. Part 5 contains an appendix of images providing additional visual detail of all the stages of the digitization and representation work, complementing the images provided throughout the narrative of the book. The results have broad implications for future efforts of this type. The team's careful and thoughtful consideration of the wide variety of challenges encountered in this endeavor, as well as the solutions, are thought provoking, highly instructive, and carefully documented in detail. Technically and mathematically demanding, this work will be of most interest to those in computer vision and computer graphics. It will prove novel and instructive to those outside of these fields, who are interested in cultural heritage preservation, digital libraries, and archives. For the nonspecialist, it provides an interesting picture of the state of the art in these areas, grounded by solving real-world problems that are understandable and significant. I recommend it. Online Computing Reviews Service