I often scan documents by using GoodNote's function. Scan docs then write down memo with Apple Pencil on GoodNotes.
Most Hiragana and Katakana are ok, but Kanji isn't. For example, printed word is "揉捏", but recognized string will be "探程" , "操探" etc. It happens even though scanned image is clear. Those characters' shape is totally different from the original string for Japanese or Kanji users.

I would be appreciate if OCR could be improved.
I also understood kanji character's shape is complicated and OCR library might be limited. So correcting OCRed string can be alternative.