Digitally Archiving Cultural Objects: The Bayon Digital Archival Project

Sunday, February 14, 2016: 1:00 PM-2:30 PM
Marshall Ballroom North (Marriott Wardman Park)
Katsushi Ikeuchi, Computer Vision Laboratory, Tokyo, Japan
We have been conducting e-Heritage Project to model cultural heritage in digital form for preservation, promotion as well as archaeological research. As a representative example of our e-Heritage project, I will explain our Digital Bayon Project to digitize, the Bayon temple, in Angkor ruin, Cambodia. The central tower of the beautiful temple is inclining year-by-year and there is a possibility that collapse may occur; it is necessary to digitize its shape, while existing.

The Bayon temple is a large and complicated structure. Many hidden areas exist, difficult to obtain data using commercially available sensors. This requires us to develop various peculiar range sensors. The Bayon temple consists of a large building with complicated surfaces, and requires many times of measurements, which ends up a half terabyte. To process such large scale measurement data, we have developed new software to process them. We have successfully modeled the Bayon temple, 150m by 150m by 40m structure, as a digital form with 1 cm resolution.

Many ruins are left in Cambodia beside the Bayon temple. We have begun a project to transfer our technique to Cambodian people by exchanging MOU with a Cambodian agency, APSARA, which has responsibility to maintain the Angkor ruins. Under this framework, we have completed a digital model of Angkor wat as a collaboration effort.

Those modeling results are utilized not only for preservation or display but also for archeological research. The Bayon temple has 173 carved smiling faces. We have run cluster analysis of these faces data and confirmed that hyper planes exist to classify all the faces into the three groups: Deva, Devata, and Ashra types. We have also conducted similarity analysis on these faces and have found that similar faces exist at proximity positions.

While our Cambodian projects emphasized to develop techniques to handle shape information, we have also conducted the Kyushu project for focusing on color aspect, how to obtain, how to display, and how to analyze precise color importation.

The northern Kyusyu has many tumuli with painted walls. Most of them are closed to public for avoiding decay caused by human breath and other microbes brought in with human. We have decided to create video contents of those painted tumuli for displaying at Kyushu National Museum. After video completion, we have conducted a couple of analyses on the data and found several new findings.

e-Heritage projects push to develop new technologies in modeling both geometric and photometric aspects. It also helps to preserve cultural heritage in digital form. Moreover, such digital data provides new understandings of cultural heritage beyond human subjective judgements.