Transcribing Your Sources with AI: Resources for Handwritten Text Recognition and Vision Language Models
Event box
Transcribing Your Sources with AI: Resources for Handwritten Text Recognition and Vision Language Models In-Person / Online
Do you work with handwritten sources that are difficult to transcribe? Are you working with a corpus that is too large to transcribe manually? Want your sources to be machine-readable?
This hybrid workshop will introduce the principles and concepts of HTR (Handwritten Text Recognition) and the different resources available, ranging from open-source to subscription-based software. The workshop will clarify what a workflow that incorporates HTR might look like depending on the types of sources you are using from ancient to modern. It will also cover the issues of cost, sustainability, and scale. You will gain a basic introduction to three software options: Transkribus, eScriptorium, and Caracal. There will be a hands-on demonstration to test out HTR on your own sources. By the end of the workshop, you will have a clear understanding of how HTR can support your research or archival project and the confidence to start designing your own HTR workflow.
What you will learn:
- HTR (Handwritten Text Recognition) fundamentals: How does HTR work and what could it do for you?
- Tool overview: Get hands-on experience with different HTR platforms
- Workflow design: Tailor your approach based on your sources, from ancient manuscripts to modern documents
- Practical considerations: Understand the trade-offs in cost and sustainability
- Live demo: Test HTR tools on your own sources and see the results in real-time
