This research seeks to address the long-standing challenges of incomplete metadata within archives and digital repositories, with the hopes of finding methods that can decrease the time involved in locating and matching images of varying archival subjects. This project builds upon previous research involving Artificial Intelligence (AI) methods in archival settings and aims to provide potential enhancements in regards to AI application within digital repositories. When it comes to creating descriptive metadata for photographs, the process is often time-consuming and relies on the evanescent expertise of the curator. This knowledge is key when it comes to developing an insightful and accessible experience for the end-user. However, correctly identifying images of individuals within photographic archives is particularly laborintensive and ultimately costly when it comes to the amount of time spent on processing a photographic collection; especially ones that have missing identifying metadata. Additionally, prominent individuals featured in collections may be known to archivists, librarians, and curators, but that resource is often lost when the institutional memory holders retire or leave. Lesser-known individuals within the repository may never be appropriately named if their identity is not quickly determined. Ultimately, in regards to preserving subject identity within repositories and archives, much of the information is heavily reliant on professional memory and achievable knowledge.
