Ὅσοι ἄνθρωποι, τοσαῦται γνῶμαι

A workshop to foster a community of practices and establish guidelines for HTR dedicated to Ancient Greek.

In recent years, there have been significant developments in the field of HTR with the rise of several general public software such as eScriptorium and Transkribus. In conjunction with technological progress, there has been a growing emphasis on the development of datasets and guidelines to standardize procedures and thereby facilitate their implementation.

For Latin script manuscripts, a critical mass of data appears to have been reached. However, non-Latin scripts often remain on the margins from this progress. Currently, researchers often work on their own, following diverse practices and standards tailored to the needs of their specific projects. This situation mirrors the early days of Latin script HTR, before the development of comprehensive guidelines that slowly gained the potential to transform the field. Similarly, the establishment of a collaborative and interdisciplinary community could facilitate significant progress in non-Latin script HTR.

HTR for Ancient Greek faces several challenges, including a limited and dispersed corpus compared to Latin scripts, which hinders the development of effective models. Another issue is the lack of consistent encoding practices for Greek paleographic features. While Unicode provides basic characters for the Greek alphabet, it does not account for all the paleographic elements (such as ligatures and abbreviations) found in Ancient Greek manuscripts. This gap has led to varying encoding practices, complicating data standardization and limiting interoperability.

To improve data reuse and collaboration, a more generalized approach is needed, with harmonized practices across projects to enable data aggregation, sharing, and pre-annotation.

Goals of this Workshop

This workshop is not intended to be a technical training session on HTR, but rather a space for reflection, collaboration, and community building. It aims to bring together researchers and practitioners from a range of disciplines, including philologists, paleographers, HTR users (across various languages), and dataset creators, to address the technical and methodological challenges that have hindered the development of effective HTR for Ancient Greek.

The primary goals are:

1. To foster the development of an interdisciplinary community of practice that combines complementary expertise to address the challenges posed by the automatic recognition of Greek documents.
2. To develop common guidelines for the transcription and encoding of ancient Greek texts, especially for the training of automatic text recognition models.

In short

Date & Time

15^th of July 2025, 09:00-12:30

Location

NOVA FCSH [tbc]

Workshop organizers

• Mathilde Verstraete (contact person)
• Maxime Guénette (contact person)
• Malamatenia Vlachou-Efstathiou
• Marianne Reboul
• Marcello Vitali-Rosati

See also the Acknowledgements

Want to join?

We welcome anyone interested in (digital) paleography or in developing guidelines for Handwritten Text Recognition (HTR) applied to Ancient Greek. This workshop will be held in hybrid format: participants may attend either in person or online (see the conference website for details). Please note that in-person attendance is limited to 30 participants.

To participate, you must register for the DH2025 conference and select the workshop when registering. Please note that the Early Bird registration period ends on May 4, 2025 and that registration closes on June 2, 2025.

If you have any questions, or if you’d like to suggest a case study (10–15 minutes – see Project #1 as an example), don’t hesitate to get in touch (write to Mathilde Verstraete or Maxime Guénette with a short description of your project). We’ll do our best to accommodate proposals within the available time, though we may not be able to include all of them.