Hello, I am getting some problems with reading text with OCR,for example when I read a number 123456789, the OCR of BluePrism get 123456?9, I had configured the font, and nothing more, any idea?
Thanks
That's the issue with all OCR, regardless of whether it's BP/Tesseract or something else. OCR is an interpretation and it can make mistakes, false-negatives and false-positives. If you use OCR in an RPA solution you have to accept the possibility of errors.
Try experimenting with the scale input parameter. That might improve the clarity of the image and give a better result.
Hello, I tried experimenting with Scale, and helps with text, but with numbers is the same, maybe the problems is because the colour of the font, I attached an example. Thanks a lot for your comment.
Strange, based on your image, I would have expected that to be no problem. Maybe try creating a loop that increases the scale from 1 to 10 in 0.1 increments and watch the output from the read stage to see if there is an optimal value. Then maybe try the same image converted to black and white to see if the colour has something to do with it - I doubt it but it's worth checking.
