We are excited to co-present with the University of Alberta Digital Scholarship Centre on some of the work we’ve been doing with OCR and vision language models!
Description: From the 1990s to the 2010s, the pace of development in text and handwritten tech was steady, but over the last 2-3 years, advancements have accelerated dramatically. Now, new tools are being released nearly every week with non-trivial improvements. Tools such as olmOCR, Chandra and Hunyan have revolutionized the field and provided scholars with expanded capabilities to extract text and structured data from images of text.
In this seminar, we will journey through the eras of Optical Character Recognition (OCR) technology, early rule-based tools to cutting-edge advancements that integrate Large Language and Vision models, demonstrating the current capabilities and outstanding potential. Methodology, best practices, and pitfalls will be presented.
We will offer brief demonstrations of a couple of scenarios, using Python scripts (which will be available on GitHub). These scripts are part of an effort to develop a flexible and up-to-date application for running these processes on a variety of platforms.
The Instructors: Chloë Farr (University of Victoria Computer Science, graduate student), Corey Davis (University of Victoria, Digital Preservation Librarian), and Peter Binkley (University of Alberta, Digital Scholarship Technologies Librarian)
Friday, February 6, 2026
Time: 12pm-1:30pm PT, 1pm-2:30pm MT
Location: Hybrid (UofA Digital Scholarship Centre; Zoom)
Click here for registration.