Beatrice Couture ; Farah Verret ; Maxime Gohier ; Dominique Deslandres - The Challenges of HTR Model Training: Feedback from the Project Donner le gout de l'archive a l'ere numerique

jdmdh:10542 - Journal of Data Mining & Digital Humanities, December 6, 2023, Historical Documents and automatic text recognition - https://doi.org/10.46298/jdmdh.10542
The Challenges of HTR Model Training: Feedback from the Project Donner le gout de l'archive a l'ere numeriqueArticle

Authors: Couture Beatrice 1; Verret Farah 1; Gohier Maxime 2; Deslandres Dominique 1

The arrival of handwriting recognition technologies offers new possibilities for research in heritage studies. However, it is now necessary to reflect on the experiences and the practices developed by research teams. Our use of the Transkribus platform since 2018 has led us to search for the most significant ways to improve the performance of our handwritten text recognition (HTR) models which are made to transcribe French handwriting dating from the 17th century. This article therefore reports on the impacts of creating transcribing protocols, using the language model at full scale and determining the best way to use base models in order to help increase the performance of HTR models. Combining all of these elements can indeed increase the performance of a single model by more than 20% (reaching a Character Error Rate below 5%). This article also discusses some challenges regarding the collaborative nature of HTR platforms such as Transkribus and the way researchers can share their data generated in the process of creating or training handwritten text recognition models.


Volume: Historical Documents and automatic text recognition
Published on: December 6, 2023
Accepted on: November 7, 2023
Submitted on: December 24, 2022
Keywords: Computer Science - Computer Vision and Pattern Recognition,Computer Science - Machine Learning
Funding:
    Source : OpenAIRE Graph
  • Funder: Social Sciences and Humanities Research Council
  • Deep Drug Discovery and Deployment; Funder: Social Sciences and Humanities Research Council; Code: PTDC/CCI-BIO/29266/2017

Consultation statistics

This page has been seen 540 times.
This article's PDF has been downloaded 207 times.