Case Study 2: Computer Vision and Cultural Heritage

The second case study is written by Catherine Nicole Coleman. This case study on computer vision applied to cultural heritage looks at critical points of intersection between research questions, the affordances of the technology and curatorial desires. The primary focus of this case study is Stanford Global Currents, a project completed in 2017 that applied computer vision techniques to medieval manuscripts. The discoveries and outcomes of that project are used as a point of departure to touch on related work at other institutions and independent work with computer vision applied to cultural heritage that has influenced how we think about search and discovery in libraries, archives, and museums.

This second case study for the AEOLIAN project is written based on interviews, project reports, conference papers, and published research. Some key terminology is defined and core concepts of computer vision that are essential to understanding the project are explained, but this is not a study of how computer vision works, nor does it address in any detail the methods or techniques applied in the Stanford Global Currents project. Global Currents took place from 2014 to 2017 and in the intervening years, visual feature extraction, which for Global Currents required custom-built algorithms, can now be done much more quickly and easily using commercially available services. Nor is the case study about computer vision as a field of study. It is about what can be learned from computational approaches to archival research that rely in some way on computer vision for information retrieval. The reason Stanford Global Currents remains an important case study today is not the technology they use, but what emerged from the researcher’s and curator’s engagement with the technology.