Patrick Durusau just pointed out to me this Austrian site: https://transkribus.eu/Transkribus/.
Transkribus is a system for transcribing manuscripts to text data. Its website allows one to upload graphic documents, and download an application to do the actual work. Evidently, after one has uploaded 100 transcribed documents, Transkribus will use those documents to train a Handwritten Text Recognition (HTR) engine which will automate further processing. One of the most important aspects of this project is that it allows and promotes collaboration. Manuscript transcription is a tedious and labor intensive process, and crowd sourcing is an excellent way to distribute the cost down to something reasonable.
Transkribus is funded by the EU tranScriptorium project. What is unclear is how long this funding will continue and how long the Digitisation and Digital Preservation Group of the University of Innsbruck will support the website and application. The danger with such projects is that after making major commitments of time and effort, the resources will dissolve and disappear. One should always ask the question, “What will we do if Transkribus goes away?”
Nevertheless, the project could be very useful in the near term. The website design is slick as well. I like it.
DISCLAIMER: I have not used the application or the website.
Leave a Reply