Impact of Image Enhancement Methods on Automatic Transcription Trainings with eScriptorium - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail (Preprint/Prepublication) Année : 2023

Impact of Image Enhancement Methods on Automatic Transcription Trainings with eScriptorium

Résumé

This study stems from the project Desenrollando el cordel (Untangling the cordel), which focuses on 19 th-century Spanish prints editing. It evaluates the impacts of image enhancement methods on the HTR transcriptions of low-quality documents, both in terms of printing and digitisation. We compare several methods (binarisation, deblur, defading) and present the results obtained during the training of models with the tool eScriptorium. We demonstrate that binarisation methods give better results than others, and that the combination of several techniques did not significantly improve the predictions of the HTR. This study shows the significance of using image enhancement methods with eScriptorium and paves the way for further experiments with larger and various corpora to help future projects design their HTR workflow.
Fichier principal
Vignette du fichier
Article_HTR_v4.pdf (1.64 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03831686 , version 1 (27-10-2022)
hal-03831686 , version 2 (14-02-2023)
hal-03831686 , version 3 (19-06-2023)
hal-03831686 , version 4 (04-07-2023)

Identifiants

  • HAL Id : hal-03831686 , version 2

Citer

Pauline Jacsont, Elina Leblanc. Impact of Image Enhancement Methods on Automatic Transcription Trainings with eScriptorium. 2023. ⟨hal-03831686v2⟩
102 Consultations
275 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More