Machines reading the archive: handwritten text recognition software

[Chris in Sussex]

Brought to my attention on Rootschat

http://blog.nationalarchives.gov.uk/blog/machines-reading-the-archive-handwritten-text-recognition-software/

I am not convinced it will ever work with all those old greasy, smudged and creased deeds I have had the pleasure of trying to transcribe at the NA but, as they say, every little helps!:)

Chris

[kiterunner]

Oh, I hope one day they develop a German version!

[Janet]

Wow.

[Merry]

Do we need entire documents transcribed? Isn't it likely common phrases will be transcribed perfectly but the names of people and places will be a hopeless bodge? Of course as researchers initially we need the names to identify likely documents before worrying about the rest of the content!

[Phoenix]

I am always amazed just how good the manual transcribing skills at TNA. A week or so back I spent half an hour or so, going through a box of chancery records, convinced the document I had wanted was mis-indexed.

But no, it wasn't. It was attached to three unrelated documents, but someone had actually spotted that it did not belong with the others and had put it into the catalogue. Many thanks to that person!