Workshop: Computer Vision & The Printed Book

This workshop by Giles Bergel was part of the DH + BH: An Interdisciplinary Conference on Digital Humanities and Book History hosted by the Digital Cultural Studies Cooperative from September 22-24, 2022.

Computer vision has made significant progress in recent years, thanks in part to developments in machine learning (or ‘AI’), and is now an eminently practical tool for the book historian. Computers can now reliably match the same printed page or illustration, or visualise variant typesettings or images. More challenging applications, such as detecting illustrations, segmenting pages into meaningful parts and classifying their content, are within reach. This workshop will introduce participants to free and open-source software tools and demos maintained by the University of Oxford’s Visual Geometry Group and developed in collaboration with book historians and others. Attendees will leave the workshop knowing how to match, differentiate, classify and annotate images of various kinds of books and prints. No previous knowledge of computer vision or coding ability is required.