Idea: implement OCR for accreditation certificate

[punchagan]

Sometimes people could enter wrong dates for the accreditation start date. WFDF didn't have an API to get the dates, but OCR could be easy to do given we know exactly what the document looks like. Some pre-processing to deal with the transparent background, followed by local OCR should do a reasonably good job.

We could do it post upload on the server, to validate the entered date, if that's easier. But, doing it on the client side using something like tesseract.js might give better UX.
https://github.com/naptha/tesseract.js/

(The issue was brought back into focus by Tara, recently)