Tools: Team One

From MarineLives
Revision as of 09:33, June 26, 2016 by ColinGreenstreet (Talk | contribs)

Jump to: navigation, search

Team One: working on semi-automated recognition of handwritten manuscripts

Team summary

We will explore line and text block recognition of legal documents using software tools developed by the Transkribus project. If we can get a Java coder on the team, we will embed the Transkribus tools in the MarineLives wiki. The team will work with C17th records from the English High Court of Admiralty and from the King's Bench. We will explore whether Transkribus tools can be used by volunteers to create metadata for virgin manuscripts for which there are neither existing metadata, keywords, nor full text transcriptions. We are interested in both the software and workflows required to systematise the creation of metadata and keywords to make previously "invisible" manuscript images discoverable.



Transkribus tool set

[ADD DATA]



Transkribus trained on High Court of Admiralty data set

Transkribus recognition engine has been trained on Admiralty Court records

[ADD DATA]