An introduction to Transkribus – automated text recognition for historical documents

Transkribus is a Research Infrastructure for automated transcription, recognition and searching in historical documents – whether they are handwritten or printed, date back to the middle-ages or represent modern writing in Latin, English, Finnish, Arabic or any other character based alphabet. Transkribus is part of the H2020 Project READ (Recognition and Enrichment of Archival Documents). For more information visit:

http://transkribus.eu/
http://read.transkribus.eu/

The webinar consists of three building blocks:
1. Demonstration. We will demonstrate the main features of Transkribus using real world examples. Significant progress has been made in the last few years in the areas of layout analysis, handwritten text recognition and keyword spotting. Participants will hear about latest developments in the field and understand the main concepts behind them.
2. Content from users. Participants are encouraged to send us their images from any kind of document they are interested in or are working with. We especially welcome handwritten documents from archives. Please send these images either directly to Transkribus: <email@transkribus.eu>, or send them via WeTransfer. In this session we will then talk about the options for how to process these images, especially in relation to your objectives and project plans.
3. Background. In the last section participants will get an overview of the Transkribus Research Infrastructure, its background and further plans.

Sesion lead

Günter Mühlberger, Universität Innsbruck

Recording