Reading ancient handwriting using an artificial intelligence platform

Reading ancient handwriting using the AI ​​platform Transkribus

Using the online platform, even non-experts can digitize and read historical manuscripts very easily. Credit: University of Innsbruck

Using artificial intelligence, computers can decode handwritten text and make it readable for everyone. The Transkribus platform, jointly developed at the University of Innsbruck, Austria, makes this technology available to researchers and the general public. An ever-growing group of people use Transkribus to research their family history. Recently, users from all over the world met in Innsbruck.

Handwriting is as individual as people. However, computers today are able to automatically recognize handwriting in a variety of languages. The Transkribus software platform, developed by the University of Innsbruck, enables this technology to scientific communityarchives and the general public. More than 90,000 users from all over the world are already using the platform to make handwritten documents readable and searchable. An ever-growing group of people are interested in family history And they began to look for their ancestors in church records, contracts, etc. historical documents.

“Searching these documents manually can be a very tedious task. Our technology now makes researching family history much easier,” said Gunter Mullberger of the Working Group on Digitization and Digital Archiving at the University of Innsbruck, Austria, and Chairman of the Board of Directors of the European cooperative READ-COOP.

Quickly search for large groups

Archives and libraries store historical documents of inestimable value. These documents take up a lot of space. For example, documents in the Austrian State Archives fill 350 kilometers of shelves. Most of these documents are only available in handwritten form and are no longer legible to many users because they are written in a script called Kurrent, an old form of German Handwriting On the basis of calligraphic writing in the late Middle Ages.

“This is where the Transkribus platform becomes easy to use, as it automatically recognizes that handwriting and makes it readable for everyone,” explains Günter Mühlberger. In addition, documents can also be searched easily. This makes searching using historical collections much easier because hundreds or thousands of documents can be searched simultaneously for family names or other terms. Permission

Credit: University of Innsbruck

He reads German, Arabic and Chinese currents

Transkribus works with neural networks. This machine learning method has the big advantage that you no longer have to manually program recognition for each type of writing. “Users teach the machine to read handwriting,” says Günter Mühlberger. “And the machine does not tire, which means it can process thousands or hundreds of thousands or millions of pages automatically. That’s what we did for the National Archives of Finland, for example, where more than two million handwritten documents dating from before the 19th century are now available to everyone “.

The technology used is completely independent of the language and the actual text or type of writing. Transkribus recognizes not only German or modern handwriting, but also medieval texts, as well as Hebrew, Arabic or Hindi handwriting. “Right now, we’re doing experiments with the ancient Chinese,” Mollberger adds.

Great help for researchers

Transkribus applications are diverse in sciences and humanities as well. For example, the classical Innsbruck philologist William Barton, who received a START prize of €1.2 million for his research with the help of Transkribus, deciphered the entries for Karl Benedikt Hase’s 19th-century diary that were thought to be missing, handwritten in ancient Greek. The valuable information contained therein should be accessible to other areas of research.

William Barton of the Department of New Latin Studies explains, “The private and confidential diary of scholar Carl Benedict Hussey contains records of nine years. The amount of texts is enormous, there are about 2,500 pages.” “I trained the machine to model Hase’s font based on 100 pages. It is now able to read all of his diaries and reliably transcribe the text.” A recent study by the University of Edinburgh revealed that more than 400 Scientific Publications It is now produced with the help of Transkribus.

Machine learning and big data unlock archives for Europe

more information:

Presented by the University of Innsbruck

the quote: Reading Ancient Handwriting Using the AI ​​Platform (2022, Oct 11) Retrieved Oct 11, 2022 from

This document is subject to copyright. Notwithstanding any fair dealing for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.

Leave a Comment