This workshop is part of the PATT project (Potentials of Advanced Text Technologies: Machine Learning Based Text Recognition) and is intended to encourage exchange between researchers, libraries and archives. We will focus on the possibilities of layout and text recognition models in general, but also specifically on how models can be shared and reused with which platforms.
Keywords include Transkribus, CATMUS-Medieval, HTR-United, hugging face, and e-scriptorium.
Programm:
From 13.15: Arrival
13.30–14.30: Ariane Pinche (CNRS, online):
First Steps in ATR for Historical Documents: Turning Manuscripts into Shareable Data.
14.50–15.10: Tobias Hodel (Universität Bern):
Building and Using Large Text Recognition Models: TrOCR & Huggingface
15.15–15.35: Christiane Sibille & Johannes Wahl (ETH):
ATR@ETH Library
15.40–15.55: Sarah Murer & Jonathan Hauller (StAZH):
From Layout to Letters: Handwritten Text Recognition and Layout Analysis in Early Modern Texts at the State Archives of Zürich
16.00–16.30: Final Discussion and Perspectives
This workshop will be held in a hybrid format. Please send registrations (physical and online) to
Alicia Schümperli (sekteuscher@hist.uzh.ch). For online participation, you will receive the link after
registration.