Hi all, I am trying to build some graphs from replays of starcraft. This is the kind of image that I use and I have croping functions that get each zone of interest (minerals, gas, workers, etc):
[image: stats.jfif] I do the following steps: - I invert with img = 255 - img (img being a numpy array) - I run image_to_string(stat_img, config='--psm 6 --oem 1') Results without inversion are really bad, whilst when using inversion results get quite better. Still I get many times letters instead of numbers (e.g 8B instead of 80 or ee instead of 0), and I feel this could be improved since the data source is so standard. What do you think I could add to this pipeline to improve results? Thanks! -- 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/5e557619-798d-4d21-ad44-ac82dc2d9747n%40googlegroups.com.