The Early Portuguese Printing OCR Model in Transkribus

on innovations in the transcription of historical printed documents and the research infrastructure for historical linguistics and digital paleography in Brazil

Authors

DOI:

https://doi.org/10.1590/1983-3652.2026.64127

Keywords:

Digital humanities, Historical linguistics, Digital paleography, Digital sovereignty, Artificial intelligence

Abstract

This paper presents the Early Portuguese Printing (EPP) Optical Character Recognition (OCR) model, developed on the Transkribus platform, and discusses the potential, the challenges and the history of such tools for research with historical documents. Transkribus, maintained by the European cooperative READ-COOP, allows researchers to train specialized AI models on the characteristics of their own corpora. The EPP model was specifically trained for the transcription of printed materials in the Portuguese language from the 16th to the 19th centuries, using a corpus of grammars and linguistic works from the period. With a training set of 142,606 words (745 pages), the EPP achieved a Character Error Rate (CER) of only 2.58%. This result represents a significant advancement, as it demonstrates the potential of such tools for creating historical quantitative corpora on large-scale and in less time, while maintaining accuracy in the transcription of diacritics, typographical symbols and Greek characters, elements that often limit the effectiveness of general-purpose OCR tools. Furthermore, the EPP is available as a base model that can be adapted by other researchers for different printed collections, through fine-tuning. However, in addition to publicizing the tool’s potential, this paper also problematizes its nature. As it belongs to a private European entity and it is a Software as a Service product, the use of Transkribus raises questions about data centralization and the sustainability of its application on large-scale Brazilian research projects, considering the future and the volume of our historical archives.

Downloads

Download data is not yet available.

References

ALKENDI, Wissam; GECHTER, Franck; HEYBERGER, Laurent; GUYEUX, Christophe. Advancements and Challenges in Handwritten Text Recognition: A Comprehensive Survey. Journal of Imaging, v. 10, n. 1, p. 1–18, 2024. DOI: 10.3390/jimaging10010018.

BARROS, João de. Grammatica da língua portuguesa com os mandamentos da santa madre igreja. Lisboa: Luís Rodrigues, 1539. Autor: João de Barros (1496–1570). Disponível em: http://acervo.bndigital.bn.br/sophia/index.asp?codigo_sophia=504. Acesso em: 4 set. 2025.

BARROS, João de. Grammatica da lingua portuguesa. Olyssipone: apud Lodouicum Rotorigiu[m], Typographum, 1540. Autor: João de Barros (1496–1570). Disponível em: https://purl.pt/12148. Acesso em: 4 set. 2025.

BAUDRY, Hervé; PEDRO, Susana Tavares. Humanidades digitais e documentos históricos: transcrever, catalogar. Cultura. Revista de História e Teoria das Ideias, Lisboa, n. 41–42, p. 235–246, 2023. Disponível em: https://revistas.fcsh.unl.pt/cultura/article/view/1223. Acesso em: 4 set. 2025.

BENGIO, Yoshua. Learning Deep Architectures for AI. Foundations and Trends in Machine Learning, Hanover, v. 2, n. 1, p. 1–127, 2009. DOI: 10.1561/2200000006.

BUESCU, Maria Leonor Carvalhão. A Gramática da Linguagem Portuguesa (Fernão de Oliveira, 1536): Introdução, leitura actualizada e notas. Lisboa: Imprensa Nacional-Casa da Moeda, 1975.

GOODFELLOW, Ian; BENGIO, Yoshua; COURVILLE, Aaron. Deep Learning. Cambridge: MIT Press, 2016.

HAUGEN, Odd Einar (ed.). MUFI Character Recommendation. 4.0. ed. Bergen: Medieval Unicode Font Initiative, 2015. Disponível em: https://nva.sikt.no/registration/0198f0c7becc-8aff5ed6-4690-4727-96bf-69e5d752c276. Acesso em: 2 jan. 2026.

LEÃO, Duarte Nunes de (fl. 1530–1608). Origem da Lingoa portuguesa. Lisboa: Pedro Crasbeeck, 1606. Disponível em: https://purl.pt/50. Acesso em: 4 set. 2025.

MACHADO, José Pedro (ed.). João de Barros: Gramática da Língua Portuguesa. 3. ed. Lisboa: Sociedade Astória, 1957.

MCENERY, Tony; HARDIE, Andrew. Corpus Linguistics: Method, Theory and Practice. Cambridge: Cambridge University Press, 2012.

NARANG, Sonika Rani; JINDAL, Munish Kumar; KUMAR, Munish. Ancient Text Recognition: A Review. Artificial Intelligence Review, v. 54, p. 1–28, 2021. DOI: 10.1007/s10462-020-09827-4.

NOCKELS, J.; GOODING, P.; TERRAS, M.; AMES, S. Understanding the Application of Handwritten Text Recognition Technology in Heritage Contexts: A Systematic Review of Transkribus in Published Research. Archival Science, v. 22, n. 3, p. 367–392, 2022. DOI: 10.1007/s10502-022-09397-0.

OLIVEIRA, Fernão de. Grammatica da lingoagem portuguesa. Lisboa: Germão Galharde, 1536. Autor: Fernão de Oliveira (c. 1507–1581). Disponível em: https://purl.pt/120. Acesso em: 4 set. 2025.

PACHECO ROCHA, Saulo Rogério. O Português das “Obras Proueitosas” e dos “Ilustres & Muito Reuerendos Senhores”: uma análise da mudança nos padrões de colocação pronominal clítica em gramáticas portuguesas do século 15 ao 19. 2025. f. 202. Dissertação (Mestrado em Linguística) – Universidade Federal de Santa Catarina, Florianópolis.

PUIGCERVER, Joan. PyLaia: Deep Learning-based Handwriting Recognition Toolkit. [S. l.: s. n.], 2017. Disponível em: https://github.com/jpuigcerver/PyLaia. Acesso em: 4 set. 2025.

PUIGCERVER, Joan. A Probabilistic Formulation of Keyword Spotting. 2018. Tese (Doutorado) – Universitat Politècnica de València, València.

READ-COOP SCE. Character Error Rate and Learning Curve. Innsbruck: [s. n.], 2025. Disponível em: https://help.transkribus.org/character-error-rate-and-learning-curve. Acesso em: 28 mar. 2026.

SPOLSKY, Joel. The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!) [S. l.: s. n.], out. 2003. Disponível em: https://www.joelonsoftware.com/2003/10/08/the-absolute-minimum-every-software-developer-absolutely-positively-must-know-about-unicode-and-character-sets-no-excuses/. Acesso em: 28 mar. 2026.

TERRAS, M. et al. The Artificial Intelligence Cooperative: READ-COOP, Transkribus, and the Benefits of Shared Community Infrastructure for Automated Text Recognition. Open Research Europe, v. 5, n. 16, 2025. DOI: 10.12688/openreseurope.18747.2.

WANG, Haifeng; PAN, Changzai; GUO, Xiao; JI, Chunlin; DENG, Ke. From Object Detection to Text Detection and Recognition: A Brief Evolution History of Optical Character Recognition. Wiley Interdisciplinary Reviews: Computational Statistics, 2021. DOI: 10.1002/wics.1547. Acesso em: 4 set. 2025.

WANG, Junmiao. A Study of the OCR Development History and Directions of Development. In: HIGHLIGHTS in Science, Engineering and Technology. [S. l.: s. n.], 2023. v. 72, p. 409. DOI: 10.54097/bm665j77. Acesso em: 4 set. 2025.

Published

2026-09-30

Data Availability Statement

Os dados de pesquisa estão disponíveis em repositório.

How to Cite

PACHECO ROCHA, Saulo Rogério. The Early Portuguese Printing OCR Model in Transkribus: on innovations in the transcription of historical printed documents and the research infrastructure for historical linguistics and digital paleography in Brazil. Texto Livre, Belo Horizonte-MG, v. 19, p. e64127, 2026. DOI: 10.1590/1983-3652.2026.64127. Disponível em: https://periodicos.ufmg.br/index.php/textolivre/article/view/64127. Acesso em: 20 nov. 2026.