Another OCR issue...

Either I don't understand OCR or this thing just isn't working. I am trying to capture a clearly visible word on the screen. I have changed the DPI, given it all valid characters, turned ignoreSpaces on and off, fine tuned the search rectangle and all I get is "LI" when I'm putting the search rectangle around the entire word.

here is the code:
put readtext ((1235, 80, 1345, 93), validCharacters:"LINKDOWN", contrast:on, contrastColor:black, DPI:dpi) into widgets

and here is the screenshot of the word needed to be read with OCR...


  • The "dpi" variable next to the DPI parameter is just a variable I'm changing in a repeat so I can figure out what the best DPI is. It starts reading it at around 60-80 dpi.
  • Think I might of just figured it out...thought the contrast was the background color, but "I" is wrong...when I changed it to White I am getting better results..
  • Hello,

    I'm glad that you are seeing better results.

    Just so you know, if you are looking to have the OCR read uppercase letters only, you should add the parameter:
    This may or may not help in this situation.

  • how do I get rid of the \n after every recognition? basically once I get it recognized it adds a \n onto which I think is a linespace right? besides parsing it after I have it, is there a way to remove that when it captures it?
Sign In or Register to comment.