DESCRIPTION
	GT4HistOCR contains ground truth for research in Optical Character 
	Recognition (OCR) technology applied to historical printings in German
	Fraktur and Early Modern Latin.

	The ground truth comes in pairs of images of single printed lines as
	they appear in book pages (*.png) and their corresponding diplomatic 
	transcriptions (*.gt.txt), which are UTF-8 strings preserving the 
	character forms (glyphs) as much as possible within the UNICODE 
	standard. These pairs of line images and their transcriptions can be 
	directly used to train recognition models with, e.g., the open source 
	OCR engines OCRopy or Tesseract.

	The subcorpora contain a certain number of shuffled lines from 
	books printed in a variety of historical fonts.

	subcorpus	      description			       # lines
	----------------------------------------------------------------------
	dta19		      first editions of 19th c. German books   243,942	
	Early Modern Latin    Latin printings 1471-1686			10,288
	KALLIMACHOS	      Narrenschiff editions 1488-1509		20,929
	Refence Corpus ENHG   German printings 1476-1499		24,766
	RIDGES		      Herbals printed 1487-1870			13,248
								--------------
								sum:   313,173

	In addition to the ground truth, there is a perl script under tools/ 
	which may be adapted to harmonize existing transcriptions of historical 
	books by different transcribers.
	
	Pretrained OCRopy recognition models for the Early Modern
	Latin Corpus (latin1 and latin2), the RIDGES corpus (ridges1 and 
	rigdes2), and the incunabula printings of the Reference Corpus ENHG
	are found under models/. For more information on these models see our
	paper cited below which also contains references to the literature.

AUTHORS
        The data have been generated in the following projects:
	
	* dta19: http://www.deutschestextarchiv.de/
	* Early Modern Latin: http://www.cis.lmu.de/ocrworkshop/
	* KALLIMACHOS: http://kallimachos.de
	* Reference Corpus Early New High German: 
		https://www.linguistics.ruhr-uni-bochum.de/ref/
	* RIDGES: https://korpling.org/ridges

	They were collected and put in the current form by
	Uwe Springmann, Christian Reul, Stefanie Dipper, and Johannes Baiter.

COPYRIGHT
	This data collection has a CC-BY-SA 4.0 license.

SEE ALSO
	If you use these data, please cite the following paper:
	
	Uwe Springmann, Christian Reul, Stefanie Dipper, Johannes Baiter (2018).
	Ground Truth for training OCR engines on historical documents in German
	Fraktur and Early Modern Latin.

LAST CHANGE
	August 2018, uvius@gmx.de