External reference-based Ocr quality check
#19
by balsab - opened
A rough quality overview for the documents in the corpus. Uses an external dictionary to count danish tokens in each page, and compares the results to other external sources of alleged "good" quality. Has problems, but is relatively quick and does not require too much resources. Included is a markdown with a more detailed rundown of the process.
balsab changed pull request status to merged