Horizon Impact Award 2020 for READ Project

  • By Claudia Vitt
  • 2020-10-08
  • Innovation
  • Excellence

The READ project implements a virtual research environment and contributes to the preservation of European heritage.

READ provides technology for the automated recognition, transcription, indexing, and enrichment of handwritten archival documents.
READ provides technology for the automated recognition, transcription, indexing, and enrichment of handwritten archival documents.
Picture: Pixabay

On the occasion of the European Research and Innovation Days taking place from 22-24 September 2020, the READ (Recognition and Enrichment of Archival Documents) project from Robert Sablatnig and his team (Markus Diem, Stefan Fiel, and Florian Kleber) from our research unit Computer Vision received the Horizon Impact Award 2020. Out of ten finalists, the Transkribus/READ project, using AI to access and analyze historical documents and archives, won the award as one of the five best projects completed in 2019/2020.  


The Style of Writing 

Whether transcribing historical documents from the middle ages or the 20th century—often handwriting is hard to read for human beings, and all the more it is for machines. Handwritten scripts are as individual as their writers. Not only do they vary from one country to another, but they are also dependent on the language, the character sets, abbreviations, and writing styles in use during a given time. The same is true for the layout of archival documents.

In READ, one can find administrative records from ministries, councils or communes; court decisions and treaties; lists of the names of immigrants and emigrants; personal letters, notebooks, diaries and postcards; cadastre books and maps; shipping records; and many other types. READ will do an automated segmentation and transcription (HTR processing), and it will also be possible to train specific scripts.  


READ: Recognition and Enrichment of Archival Documents

The overall objective of READ is to implement a Virtual Research Environment. Archivists, humanities scholars, computer scientists, and volunteers can collaborate with the ultimate goal of boosting research, innovation, development, and usage of cutting edge technology for the automated recognition, transcription, indexing, and enrichment of handwritten archival documents. 


 This Virtual Research Environment will benefit from research, tools, data, and resources generated in multiple national and EU funded research and development projects. It will provide a basis for sustaining the network and technology in the future. This ICT-based e-infrastructure will address the Societal Challenge mentioned in Europe in a Changing World, namely the transmission of European cultural heritage and the uses of the past as one of the core requirements of a reflective society.

Based on research and innovation enabled by the READ Virtual Research Environment, one will explore and access hundreds of kilometers of archival documents via the full-text search. It is now possible to open up one of Europe’s rich cultural heritage’s last hidden treasures. One of the main ambitions of READ is to revolutionize access to handwritten document collections. 


A Tent for Scanning

People dealing with deciphering historical letters, postcards, manuscripts, or medieval documents can get help: The ScanTent was developed at TU Wien Informatics and is the optimal solution for scanning lose or bound documents on the go. This comes along with low cost and high quality.