Let the user provide possible character sets for OCR
Opened this issue · 0 comments
Avasam commented
Is your feature request related to a problem? Please describe.
- We could let the user provide a set of "characters that could appear on screen", to give tesseract as an allow-list, improving consistency and speed.
- This may be the same thing, but similarly for the language to match against.
- We could validate that user-provided characters are all supported by tesseract.
Additional context
#272 (review)
CC @ston1th & @realRammbob