> Why? Would it be more cost effective to just have it double/triple-keyed and compare the transcriptions?
Indeed. took me 30 seconds to find it (and that was just from typing in the first line and deciding if it was a 1 or an l ....) https://www.google.com/search?q=IyEvYmluL2Jhc2gKCiMgQ29uZ3Jhd https://shogo82148.github.io/ But there may also be much more efficient ways to crack this particular nut than using OCR. Such as a little searching on the web: https://github.com/energelpen/UNIQLO_Akamai_T-Shirt_Bash Thanks for the reply. Yes, I did find these existing transcriptions. I was more interested in understanding how to get this to work with an OCR pipeline. > You don't say what pre-processing you did. Did you remove all the orange? Anything else? I provide the command line I used and the image to reproduce the issue. As shown there, no pre-processing yet. > That's a basic shell quoting issue that the documentation for your shell should cover. Thanks. I thought so, I was just flagging that I knew this wasn't yet covered in te command line I provided. Are there better settings to use for such a use case? > Maybe? But there may also be much more efficient ways to crack this particular nut than using OCR. That's what I am interested in. If not OCR or human transcription, what would you suggest? Best wishes, Tom -- 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 visit https://groups.google.com/d/msgid/tesseract-ocr/f50f4cfb-5a55-4d74-8a7b-c7570fe52b9an%40googlegroups.com.
