Searchable Digitized Manuscript Collections: An Opportunity to Read Serbian Cyrillic

Abstract

The READ (Recognition and Enrichment of Archival Documents) project has the potential to revolutionise access to historical collections held by cultural institutions all over Europe. This project was implemented in the period 2016/2019. It was funded by the European Commission, and involved 13 partners from the European Union. The overall objective of READ was to build a virtual research environment where archivists, humanities scholars, IT specialists and volunteers would 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.
Since its launch in 2016, in line with its concept of creating virtual research environment, the READ project was developing advanced text recognition technology on the basis of artificial neural networks. Research in pattern recognition, computer vision, document image analysis, language modelling, but also in digital humanities, archival research and related fields has seen unprecedented progress in recent years, and European research groups are on the forefront of this specific field. Newly developed technologies and tools are integrated via publicly available infrastructure – the Transkribus platform.
The primary goal of Transkribus is to support users who transcribe printed or handwritten documents. Only a few years ago, it was still in the realm of fantasy that computers would become able to read historical scripts and to automatically recognise and transcribe the handwritten text of documents from the past centuries. On the other hand, users of Transkribus are able to extract data from handwritten and printed texts via HTR (Handwritten Text Recognition) technology and search digitized text without retyping, using sophisticated technology known as KWS (Keyword Spotting), while simultaneously contributing to the improvement of the same technology thanks to machine learning principles. The automated recognition of a wide variety of historical texts has significant implications for the accessibility of the written records of global cultural heritage.