PSNC Digital Libraries Team

Activities performed by PSNC Digital Libraries Team in frame of the IMPACT project resulted in a set of full text versions of selected Polish historical documents from four digital libraries in Poland:

• Elbląska Library (Elbląska Digital Library)
• The Kórnik Library of the Polish Academy of Sciences (Digital Library of Wielkopolska)
• Poznań University Library (Digital Library of Wielkopolska)
• The Institute of Journalism, University of Warsaw (Digital Library of Polish and Poland-Related News Pamphlets)
• Wrocław University of Environmental and Life Sciences (Dolnośląska Biblioteka Cyfrowa)

Below you can find a list of the documents with corresponding source data (master files) and ground truth data (full text versions). Full text versions are enconded in PAGE XML format. The description of this format can be found here. Full text documents have accuracy around 99.95%. There are two versions of full text:

1. Full text versions on region level (paragraph), where accuracy is around 99.95%.
2. Full text versions on region level (paragraph), with additional information about coordinates for lines, words and characters. This additional information comes from the optical character recognition process and was not corrected/changed in any way. Nevertheless, this kind of information can be helpful in case one wants to have approximate coordinates of the word or character on the image. Accuracy level for region level is still around 99.95%.

All above resources are also available in the Poliqarp search engine, hosted by the Formal Linguistic Department of the University of Warsaw and available at http://poliqarp.wbl.klf.uw.edu.pl/en/.

Moreover for 478 files from all the files processed in frame of the IMPACT project, a detailed information related to coordinates on glyph level has been prepared. These files are available at:

• Antiqua files:
  • Full text versions on region, line, word and glyph level.
  • Master files.
• Gothic (fraktur) files:
  • Full text versions on region, line, word and glyph level.
  • Master files.

Above detailed data has been used in the experiment related to comparison of Tesseract and FineReader OCR engines. The report from this comparison is available here, and the trained datasets are available for download here.

Altogether 4693 files were processed, corresponding full text versions have 6890677 characters. Size of the master files is around 16,5GB. Size of the full text is around 300MB, and size of full text with additional information is 700MB.

Master files are available thanks to particular libraries. All materials are available under Creative Commons Attribution 3.0 Unported License.

Aletheia Sans font can be used to view characters in the ground truth documents and it is available for download here. Aletheia Sans is an extension of the font DejaVu Sans and as such it is adding numerous historical characters to one of the most complete and free Unicode fonts. It was created and is maintained by the PRImA Research Group (www.primaresearch.org/tools) as part of the ground truth production tool Aletheia which is described in “Aletheia – An Advanced Document Layout and Text Ground-Truthing System for Production Environments”, C. Clausner, S. Pletschacher, A. Antonacopoulos, Proceedings of the 11th International Conference on Document Analysis and Recognition (ICDAR2011), Beijing, China, September 2011, pp. 48-52. This font is available under DejaVu Seans license (availble for download here).

In case of questions or remarks related to the data, please contact us.