Ligorio 0.3

Free Public AI Model for Handwritten Text Recognition with Transkribus

The Ligorio 0.3 model is based on the handwriting of the Italian antiquarian Pirro Ligorio (Naples
1513 – Ferrara 1583).

The training data consists of 57.003 words taken from volume 15, book XVII, letter R, of the
alphabetically organised Enciclopedia del Mondo Antico. This Encyclopaedia, written between
1568 and 1583, is now in the State Archives in Turin and can be accessed online on their Digital
Library: [volume

The languages used in the text are Italian, Latin and Greek.
The transcription criteria used to form the HTR model are:
▪ Normalisation of «V» to «U» except in Latin inscriptions;
▪ Preservation of the diacritical marks and punctuation as used by the author except for the
parts in Greek;
▪ Transcription, according to the grammatical rules of the Italian language, when the use of
capital and small caps are not distinguishable.

Due to an error that occurs when visualizing the Roman sestertius and denarius in Latin
inscriptions, used by Pirro Ligorio, the symbols were replaced with others that the author would
have never used in his text. The model was trained accordingly and the symbols were chosen from
the Astronomical group already present on the virtual keyboard in Transkribus:
Roman sestertius sign (U+10198) ➛ ☿ (U+263F Mercury)
Roman denarius sign (U+10196) ➛ ♀(U+2640 Female sign).

This model was created as part of a Phd project and is part of a wider project named Ligorio
Digitale. For more information regarding the background of the model please visit: [Italian].

Citation information (CMSO, website content):
Agostini, Giorgia. “The Ligorio 0.3 model, Free Public AI Model for Handwritten Text Recognition
with Transkribus”. Last modified [date]. [model link]

Model Overview

Ligorio 0.3 PyL
Giorgia Agostini
Model ID:
Greek, Italian, Latin
Latin alphabet
CER on validation set:
2.70 %
Simply upload a picture and test this model

By uploading an image, you accept our terms and conditions and our privacy policy

Ligorio 0.3 PyL is freely available to everyone

Get started with Transkribus and use it for your own Material
You can use this model to automatically transcribe Handwritten 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 Latin alphabet documents into editable and searchable text.