I still fail to understand why Tesseract performs so poorly. Isn’t it made for OCR in screenshots? Doesn’t it understand Russian at all?
On Monday, November 2, 2020 at 5:45:41 PM UTC+1 d-ka wrote: > Well, that’d require much additional logic because the general layout > entails quite a diverse segmentation. > > The main question is, why Tesseract obviously has severe trouble with > clear Russian, no-noise PNGs—and what could be done about it. > > On Thursday, October 8, 2020 at 7:08:28 AM UTC+2 shree wrote: > >> Give each region of interest separately. >> >> >> <http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> >> Virus-free. >> www.avg.com >> <http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> >> >> <#m_-4931165716036561972_m_-7139881135647065081_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2> >> >> On Wed, Oct 7, 2020 at 6:01 PM 'd-ka' via tesseract-ocr < >> tesser...@googlegroups.com> wrote: >> >>> >>> I’d like to process Duolingo screenshots with Tesseract, in order to >>> have exercises worth reiterating in a searchable form (i.e. a text file). >>> However, it just yields gibberish: >>> >>> > tesseract.exe img.jpg img.jpg -l rus+eng --tessdata-dir "\tessdata" >>> >>> [image: FXjEk.png] >>> >>> Э 20:22 >>> 51МАВО\М/ >>> Тгапз(а{е {15 5еп{епсе >>> Апу диес00п5 >>> Уоч аге согтес& |" >>> СОМТИМЧЕ >>> Ч 4 >>> >>> >>> - For my inherent neural network, it’s easy to resolve: clear >>> contrasts, easy font, no scanning artifacts. >>> - It doesn’t read the actual Russian part at all (Вопросы есть?), >>> yet I don’t find the font weight too light or thin. >>> - No luck with greyscale or increased contrast, or by varations of >>> rus+eng. >>> - I assume that it’s implicitly UTF-8 >>> >>> <https://stackoverflow.com/questions/9976592/tesseract-does-not-recognize-russian> >>> >>> and that I already have appropriate trained data >>> >>> <https://stackoverflow.com/questions/63431711/easily-readable-text-not-recognized-by-tesseract> >>> . >>> - What could help Tesseract to properly parse this seemingly easy >>> imagery? >>> >>> Thanks so much! >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "tesseract-ocr" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to tesseract-oc...@googlegroups.com. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/tesseract-ocr/4978d94a-ec7d-4bce-b8be-cd58576d4ab2n%40googlegroups.com >>> >>> <https://groups.google.com/d/msgid/tesseract-ocr/4978d94a-ec7d-4bce-b8be-cd58576d4ab2n%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> >> >> >> -- >> >> ____________________________________________________________ >> भजन - कीर्तन - आरती @ http://bhajans.ramparivar.com >> > -- You received this message because you are subscribed to the Google Groups "tesseract-ocr" group. To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-ocr+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/18aaa2c8-a580-4932-9cc9-8659fe27a1a0n%40googlegroups.com.
