Exploring Data Provenance in Handwritten Text Recognition Infrastructure: Sharing and Reusing Ground Truth Data, Referencing Models, and Acknowledging Contributions. Starting the Conversation on How We Could Get It Done

• DOI: 10.46298/jdmdh.10403
• Publication Date: 2024-03-18T15:05:16Z
• Authors: C. Annemieke Romein; Tobias Hodel; Femke Gordijn; Joris J. van Zundert; Alix Chagué; Milan van Lange; Helle Strandgaard...; Andy Stauder; Jake Purcell; Melissa M. Terras; Pauline van den H...; Carlijn Keijzer; Achim Rabus; Chantal Sitaram; Aakriti Bhatia; Katrien Depuydt; Mary Aderonke Afo...; Anastasiia Anikina; Elisa Bastianello; Lukas Vincent Ben...; Arno Bosse; David Brown; Ash Charlton; André Nilsson Dan...; Klaas van Gelder; Sabine C.P.J. Go; Marcus J.C. Goh; Silvia Gstrein; Sewa Hasan; Stefan von der Heide; Maximilian Hinder...; Dorothee Huff; Ineke Huysman; Ali Idris; Liesbeth Keijzer; Simon Kemper; Sanne Koenders; Erika Kuijpers; Lisette Rønsig La...; Sven Lepa; Tommy O. Link; Annelies van Nispen; Joe Nockels; Laura M. van Noort; Joost Johannes Oo...; Vivien Popken; María Estrella Pu...; Joosep J. Puusaag; Ahmed Sheta; Lex Stoop; Ebba Strutzenbladh; Nicoline van der ...; Jan Paul van der ...; Barry Benaissa Trouw; Geertrui Van Synghel; Vladimir Vučković; Heleen Wilbrink; Sonia Weiss; David Joseph Wrisley; Riet Zweistra
• Keywords: Ground truth

This paper discusses best practices for sharing and reusing Ground Truth in Handwritten Text Recognition infrastructures, as well ways to reference acknowledge contributions the creation enrichment of data within these systems. We discuss how one can place a repository and, subsequently, inform others through HTR-United. Furthermore, we want suggest appropriate citation methods ATR data, models, made by volunteers. Moreover, when using digitised sources (digital facsimiles), it becomes increasingly important distinguish between physical object digital collection. These topics all relate proper acknowledgement labour put into digitising, transcribing, HTR data. also points broader issues surrounding use machine learning archival library contexts, community should begin record both provenance.