Developing an Image Database for IFA's Aphrodisias Project

For more than 40 years, teams of archeologists have been excavating and documenting the archeological site at Aphrodisias, Turkey. The Institute of Fine Arts (IFA) at NYU has been sending teams to participate in this activity since 1961. As a result, IFA has acquired thousands of slides, drawings, and descriptions of artifacts related to these excavations, which are presently stored (in analog form) at the Institute of Fine Arts on 77th Street.

The Aphrodisias Project at NYU is currently under the supervision of Christopher Ratté, Associate Professor of Classics and Fine Arts. In order to provide easy and interactive access to these materials, Prof. Ratté developed a series of databases devoted to information about the different categories of artifacts and records that relate to the site: Sculpture, Bibliography, Coins, Contexts, Drawings, Inscriptions, New Objects, Photos, and Slides. These databases were developed using FileMaker software.

In the summer of 2002, the Humanities Computing Group (HCG) at NYU's Information Technology Services (ITS) began consulting with the IFA Aphrodisias team on two separate but related projects: the development of online access to some of these existing datasets and the digitization of the extensive slide collection of Aphrodisias materials. The existing FileMaker databases were mounted on a web server, and a data entry, search, and retrieval interface was then developed using Lasso (a middleware that is typically used with FileMaker databases).

Working with Prof. Ratté and his research assistants, led by Leah Long, the HCG developed a workflow for scanning the slides, and some guidelines for scanning, benchmarking and saving the results. We based the guidelines on existing standards and best practices for image digitization, such as those documented in the NINCH Guide to Good Practice for the Digital Representation and Management of Cultural Heritage Materials and other standards-based works (see http://www.nyu.edu/its/humanities/ninchguide/). We also developed metadata for the digital images, based on the Dublin Core Guidelines (see http://www.dublincore.org for more information).

Consequently, the project has two components—digitizing the slides, and then entering metadata about the digital images into a new database application, which was designed using PHP/MySQL for storage of the metadata. The image scanning takes place at the Studio for Digital Projects and Research, an ITS/NYU Libraries collaborative facility based in the Bobst Library, using a Nikon Super Coolscan 4000 ED slide scanner. The people responsible for scanning follow a strict workflow to ensure the quality of the scans—that they are taken at the highest possible resolution and, perhaps most importantly, in a consistent fashion. This has made it easier to work on the project incrementally; more digitization can be carried out as time and/or funding becomes available. Everyone working on the project has access to the settings and workflow that have been used. Thus, no matter when or where the work takes place, the results will be consistent.

The workflow includes the following steps:

• Assess the slides that are to be scanned, check the condition of the slide, and catalog the details.
• Take note of the slide number.
• Clean dust from the slide and the scanning surface.
• Preview the slide using the scanner software.
• Crop the image.
• Scan the cropped area at 3000 DPI and a bit depth of 16.
• Save the image as an uncompressed TIFF file, with a filename identical to the slide number and a .tif extension.
• Check the digital image for quality control.
• Save the image to a folder named for the date(s) of scanning, e.g., a folder containing files scanned from 2-4 July 2002 would be named 20020602-04. This folder is then copied to CD-ROM.
• Convert the file to a JPEG web version using the automated Photoshop actions created by the HCG.

After scanning an image, metadata is entered into the MySQL database. The metadata with which we are concerned deals with the digital image and not the original archeological object. Such metadata includes the date the image was scanned, who scanned the image, which scanner was used, the dimensions of the scan, the file size in bytes, etc. The HCG has developed a simple web interface using PHP, through which the data can be entered:

Figure #1 - PHP Web Interface

The MySQL database actually records more data than is shown in the above figure, but since much of the information is constant for each image (resolution, bit depth), and taken from the system (time and date), it is included in variables that the user does not normally see.

A full record in the database looks like this:

Figure #2 - Full MySQL Record

Future stages of the Aphrodisias Project may include the development of a larger digital image collection, and the creation of online versions of all the Aphrodisias databases. We also hope to migrate the FileMaker/Lasso portion of the Project into PHP/MySQL. There is also a potential international component: another Aphrodisias project is based at King's College, University of London. Staff at the Centre for Humanities Computing at King's have been working on an XML edition of some inscriptions (see http://www.kcl.ac.uk/humanities/cch/epapp/epapp.htm). The HCG has been exchanging information with them about our work, and we are exploring the possibility of developing some collaborative projects with the London group in the future.

For more information about the Aphrodisias Project at NYU, please visit: http://www.nyu.edu/projects/aphrodisias/.

Author Biographies

Matthew Zimmerman is a Humanities Computing Specialist working in ITS Academic Computing Services. He can be reached at: matthew.zimmerman@nyu.edu.

Lorna Hughes is Assistant Director for Humanities Computing in ITS Academic Computing Services. She can be reached at: lorna.hughes@nyu.edu.

Christopher Ratté is Associate Professor of Classics and Fine Arts in Faculty of Arts and Science, and Co-Director of Excavations at Aphrodisias, Turkey. He can be reached at: christopher.ratte@nyu.edu.