Columbia University (Columbia University Media Center for Art History and Archeology)

A team from the Media Center for Art History at Columbia University will apply image processing, deep learning, and optical character recognition (OCR) techniques to investigate methods for automated processing and data extraction from digitized 35mm slides to support archival discovery. The 35mm slide library, once the cornerstone of teaching art history at all levels, now languishes in storage, with valuable image resources trapped inside its drawers. This project aims to make these legacy collections accessible again, with the potential to increase available teaching resources, make significant images available for study, and to preserve the institutional memory and pedagogical history retained in these collections. They hope to discover techniques that will allow for less time-consuming and more cost-effective processing of legacy slide collections, enabling libraries across the country to partially automate the cataloging and identification of significant images.