Solved the issue by myself. Turns out it was a derivative of one of the very first solution I tried to applied: turning the image into a gray scale. Apparently all I needed was to instead convert it to a binary image rather than a grayscale bitmap. Everything's working much better now. Thank you all! On Friday, 17 September 2021 at 18:41:45 UTC+3 AdamTuby wrote:
> [image: Screenshot from 2021-09-17 18-28-29.png][image: Screenshot from > 2021-09-17 18-28-19.png][image: Screenshot from 2021-09-17 18-28-13.png]I'm > trying to detect a paragraph in some image. I do that by preparing in > advance the very first and very last expressions in the paragraph, in order > to box all of the paragraph, tho I'm having issues with detecting the very > last expressions in the paragraph. It's detecting the characters of the > very last words, but for some reason it just returns completely different > and tho coherent (consistently the same) characters that don't match what's > actually written there. > I'm adding some png's for visual representation of the issue. Btw, excuse > me in advance if the contents of the paragraph of the text have > inappropriate contexts - I'm simply doing my testing on this specific png. > > The png's: -- 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/54459ba8-0d20-4232-b7df-1a585d461183n%40googlegroups.com.
