About the topic - tesseract is bad if the lines are curved or tilted, you should preprocess the image to adjust the lines.
Anyway, as far as a see you probably have recognition problem because the rectangles are too tight, if you put those rectangles with code you probably found the contours around the characters and called boundingRect() on them. This is ok, just after you cut out the symbols add 3-4 rows and columns of white pixels on each side and then give the expanded image to tesseract. The problem is that when a character is cut too close and it touches the border of the image recognition drops significantly and just adding some "empty" space around the character or word the recognition % is way better. -- 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 [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/tesseract-ocr. To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/e998ea0f-a088-496b-a501-7aaa1f37bfbb%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
