AMICO: The Art of Managing Art

PARTICIPANTS
Reagan Moore, Bertram Ludäscher,
Arcot Rajasekar, Amarnath Gupta,
Wayne Schroeder, SDSC

he Art Museum Image Consortium (AMICO) is a nonprofit consortium of art-collecting institutions working to enable educational use of their digital multimedia resources--currently text and images, soon audio and video sequences as well. The initial launch of the AMICO Library includes multimedia documentation for more than 50,000 works of art--from prehistoric goddess figures to contemporary paintings. Drawn from the collections of leading museums across North America, the AMICO Library is available to universities, colleges, schools, and public libraries under license for educational use. But how can this wealth of data records, of wildly different types and contents, be managed and accessed? In the AMICO University Testbed, the California Digital Library (CDL) and SDSC are working together on a technology transfer pilot project to develop state-of-the-art digital library infrastructure and services.

AMICO (Italian for "friend," pronounced like amigo) is an open, voluntary collaboration. It currently comprises 28 institutions, including the Library of Congress, the National Gallery of Art, the J. Paul Getty Museum, the Los Angeles County Museum of Art, the New York Metropolitan Museum of Art, the Art Institute of Chicago, the Boston Museum of Fine Arts, the National Gallery of Canada, and the Whitney Museum of American Art, among others.

STATE-OF-THE-ART ART

TESTBED FOR TECH TRANSFER

Figure 1. A Sample AMICO Library Data Record

This sample record for Michelangelo's chalk sketch of a nude for the Sistine Chapel ceiling contains several of the various types of data that accompany a work of art in the collection.

STATE-OF-THE-ART ART

Member institutions contribute documentation for at least 500 works of art from their collections each year, and AMICO then puts this information into a standardized database. A library entry minimally includes a descriptive "cataloging" record for a work, a digital image, and metadata documenting that image. In addition, entries may include closeups and alternate views, provenance, collection data, curators' records, original scholarly research, or related educational material (Figure 1).

"Managing this ever-growing collection of data--and metadata--is a nontrivial problem," said SDSC's Chaitanya Baru. "Fortunately NPACI and the CDL are developing technologies that complement AMICO's needs, and we have the opportunity to develop and refine our tools with a useful test case."

Baru is project manager of the Large Image Database Management phase of the AMICO University Testbed Project, which provides an opportunity for the CDL and SDSC to work together on a technology transfer pilot project that meets immediate needs of both organizations in developing advanced digital library infrastructure and services. Related efforts within the testbed include development of user tools and services and an evaluation of their efficacy, led by Susan Jurist of UC San Diego; consideration of licensing, authentication, and authorization issues, led by Ken Weiss of the CDL; and creation of an image metadata union catalog that merges the disparate image collections together, led by Weiss and Karen Coyle, also of the CDL.

TESTBED FOR TECH TRANSFER

AMICO supports research and technology deployment through development of content management and archiving technologies and the application of new image database access interfaces. SDSC is developing database and archival technologies to federate data collections and to search and display images from online databases and archival storage. These include the MIX information mediation system and the SDSC Storage Resource Broker and Metadata Catalog (SRB/MCAT) to access distributed and remote archives.

CDL standards require that all archival images include a minimum set of metadata, described in the CDL Image Metadata Document Type Definition (DTD). The AMICO effort at SDSC already has created a DTD for the AMICO metadata and has populated a relational database with these records. The MIX system uses these DTDs as a way to integrate views across diverse datasets that may have different interfaces. "Once we have received the full collection, we believe we will be able to provide full query support across all AMICO metadata," Baru said.

The use of XML and XSL style sheet scripting allows developers to create custom data-presentation interfaces for different user communities simply by changing style sheets. The AMICO project staff were intrigued by the prospect of being able to easily generate alternate presentations for the collection.

The experience gained from the testbed will help determine an architecture for CDL image databases and create a framework for CDC and SDSC collaboration. "The AMICO project provides a mechanism to test new digital library technologies for image management and access as well as data storage and archiving of image collections," Baru said. "This work is important to building the new tools and services needed to archive and access other large collections, such as the CDL's Online Archive of California." --MG *