The editions of Arnao Guillén de Brocar of BECLaR project transcribed with the help of Transkribus and OCR4all: the creation of a model for the neural network system and the exploitation of results obtained

Authors

DOI:

https://doi.org/10.13136/2284-2667/1102

Keywords:

early printed books OCR, Recurrent Neural Networks, Arnao Guillén de Brocar, Transkribus, OCR4all

Abstract

Among the software available for the recognition of old printed texts, I have decided to use two different tools, Transkribus and OCR4all, for the diplomatic transcription of the Arnao Guillén de Brocar’s editions. Firstly, the following research wants to point out the typographic and editorial characteristics of the editions of Latin classics printed by Arnao that are outstanding for the creation of a training model for the neural networks system on which Transkribus and OCR4all are based. Secondly, it will be intended to present some tools and methods to improve transcription results. Actual outcomes deserve to be shared, though the work is still at an early stage.

Downloads

Published

2022-06-29