Hi Zdenko,
Thanks for your recommendation about image format and letsgodigital trainidata. Yes, you are right. I got the digits from a segment display. I would try the training process before that i wanted to try other options. I suppose you have used the lets.traindata <https://github.com/adrianlazaro8/Tesseract_sevenSegmentsLetsGoDigital/blob/master/Trained%20data/lets.traineddata> after renaming, when i tried the same command with same psm, on the PNG image, I got .4 instead. By the way, Did you apply any processing on the image?, the edges look slightly different. tesseract 14.png out -l lets --oem 0 --psm 7 .4 Thanks for your time. On Friday, June 28, 2024 at 3:31:15 PM UTC+2 zdenop wrote: > First of all, using jpg as a format for image processing and OCR is not > very smart. > > Next: it does not seem like a very standard font... maybe you will need > to train tesseract for it. > For me, it looks like a heavy preprocessed 7-segment font... so I tried > this: > > tesseract 14.png - --psm 7 --oem 0 -l letsgodigital > 14 > > Zdenko > > > pi 28. 6. 2024 o 14:09 'uday kaipa' via tesseract-ocr < > tesser...@googlegroups.com> napísal(a): > >> I have resized the image so that text height would be around 30pxs and i >> have tried with 10px boarder as recommended in some threads here. >> I converted image to binary, and tried all PSM modes. >> I am not sure why it is not OCR'ed properly. >> >> Any help is appreciated. :) >> >> >> >> >> >> On Thursday, June 27, 2024 at 6:24:36 PM UTC+2 uday kaipa wrote: >> >>> Hi, >>> >>> I have an image having number 96 in it.(that might contains a number >>> between 0 and 100.) PFA. >>> I have used tesseract PSM from 6 to 13 and image size and font and >>> everything looks good to me. Text is recognized as 36. >>> When i try to adjust padding or other pre-processing, it would work for >>> this image and some images are recognized incorrectly. >>> >>> Can anyone recommend any other pre-processing that might improve the >>> recognition. >>> >>> *t**esseract --oem 1 --psm 7 -c tessedit_char_whitelist=0123456789.: >>> C:/Users/xxx/Desktop/test_folder/IMG_2303_2cfac/subboxes/Image_BHU32_1_PREPROCESSED_27-06-2024_17h39m53s.JPG >>> >>> new hocr* >>> >>> >>> *Many thanks in advance.* >>> >>> >>> *Regards* >>> *Uday* >>> >>> >>> -- >> 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-oc...@googlegroups.com. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/tesseract-ocr/d59827e4-6973-45af-92c0-e2aebbd7f2e7n%40googlegroups.com >> >> <https://groups.google.com/d/msgid/tesseract-ocr/d59827e4-6973-45af-92c0-e2aebbd7f2e7n%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> > -- 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/09a5c5e1-2cc7-49c2-9833-e2dc5c770203n%40googlegroups.com.
