Spanish Gothic 15th-16th Century

Free Public AI Model for Handwritten Text Recognition with Transkribus

Spanish Gothic 15th-16th Century

This general model is based on different Spanish printed books in Gothic typeface and published between 15th-16th Century. Digitized images used to create the model comes from different online repositories, mainly Biblioteca Digital Hispánica, the official repository of Biblioteca Nacional de España (BNE).

In order to create a consistent and varied extended model, around 20 pages per book were taken and several accomplished scholars participated to the transcription phase. The initial dataset is therefore based on 16 books, corresponding to 16˙816 lines and 150˙137 words; nevertheless, future users can add their own transcriptions and be included as collaborators to the next versions of the model.

For further information about the general model, the basic dataset and transcription criteria, visit the following page:

https://github.com/stefanobazzaco/HTR-model-SpanishGothic_XV-XVI_extended

The model will be freely available inside Transkribus platform, in the public model section. Its private usage is Open Access but its handling for distribution/publication requires to give appropriate credits to authors and publisher.

The dataset is available at the following page:

https://zenodo.org/record/4888927#.YLX6xqgzY2w

Creative Commons License CC BY-NC-ND 4.0 (Attribution-NonCommercial-NoDerivatives 4.0 International).

To get a pre-trained P2PaLA model for double column pages, send an e-mail to: stefano.bazzaco.1@gmail.com

To the creation of the Spanish Gothic 15th-16th Century HTR+ model took part the following projects:

  • Progetto Mambrino – Università degli Studi di Verona

resp. Anna Bognolo, Stefano Neri

https://www.mambrino.it/it

  • BIDISO (Biblioteca Digital Siglo de Oro) – Universidade da Coruña

resp. Sagrario López Poza, Nieves Pena Sueiro

https://www.bidiso.es/index.htm

  • COMEDIC (Catálogo de obras medievales impresas en castellano) – Universidad de Zaragoza

resp. María Jesús Lacarra

https://comedic.unizar.es/

  • Progetto di Eccellenza: Le Digital Humanities applicate alle lingue e letterature straniere – Università degli Studi di Verona

resp. Paolo Frassi

https://www.dlls.univr.it/?ent=progetto&id=5327

Model Overview

Name:
SpanishGothic_XV-XVI_extended (v1.0.0)
Creator:
Stefano Bazzaco
Model ID:
33106
Century:
15th, 16th
Languages:
Spanish
Script:
Gothic Script
Engine:
HTR+
Material:
Print
CER on validation set:
0.92 %

SpanishGothic_XV-XVI_extended (v1.0.0) is freely available to everyone

Get started with Transkribus and use it for your own Material
You can use this model to automatically transcribe Print documents with Handwritten Text Recgnition in Transkribus. This model can be used in the Transkribus Expert Client as well as in Transkribus Lite.
This AI model was trained to automatically convert text from images of historical Gothic Script documents into editable and searchable text.