| title | author | date | tags | ||
|---|---|---|---|---|---|
OpenArabicPE: OCR training data |
Till Grallert |
2019-05-22 |
|
This repository holds training data for OCR/HTR algorithms and some tools for converting TEI files into one .txt file per page required for this task.
The oXygen project OpenArabicPE_OCR.xpr is configured with a single transformation scenario, which makes use of an XSLT stylesheet (convert_tei-to-plain-text-pages.xsl) from another GitHub repository (convert_tei-to-markdown). In order for this conversion to work, both repositories (ocr_training-data and convert_tei-to-markdown) need to be children of the same folder and input TEI should validate against the OpenArabicPE ODD.