Need help for this OCR processing
96 views
Skip to first unread message
Eggo chen
Nov 23, 2016, 4:12:56 PM11/23/16
to tesseract-ocr
Hi All,
I am new to tesseract and need to process a pdf file using tesseract. I convert the pdf file into png (shw in below)and run through tesseract but the result is not perfect. How can I make this to work? Thank you very much in advance.
I am using Tesseract Open Source OCR Engine v3.04.00 with Leptonica.
I am new to tesseract and need to process a pdf file using tesseract. I convert the pdf file into png (shw in below)and run through tesseract but the result is not perfect. How can I make this to work? Thank you very much in advance.
I am using Tesseract Open Source OCR Engine v3.04.00 with Leptonica.
Reinaldo Crespo
Nov 23, 2016, 6:10:44 PM11/23/16
to tesseract-ocr
Have you tried different page segmentation modes?
Execute tesseract from command prompt and take a took at -psm parameter options. Report here what works best.
Eggo chen
Nov 28, 2016, 9:20:52 AM11/28/16
to tesseract-ocr
I used -psm with option from 0 through 5 and only 1, 3, 4 and 5 produce output file. With option 1 and 3 produce the most readable outcome as following.
KFS :Procuflment Card Imps://kfs-prod.adminspps.comell.=d|l/kfs/finxncizl.1’mcmment€ard.d .
mm," 1W m Hm
Pmnmmt cml IE
mam as am: mammal/241mm
explm .. whpu .u
, ' mama ma
Dou-lull m V m N
on"... m...
- mm: mmummmmmlm \
KFS :Procuflment Card Imps://kfs-prod.adminspps.comell.=d|l/kfs/finxncizl.1’mcmment€ard.d .
mm," 1W m Hm
Pmnmmt cml IE
mam as am: mammal/241mm
explm .. whpu .u
, ' mama ma
Dou-lull m V m N
on"... m...
- mm: mmummmmmlm \
Reply all
Reply to author
Forward
0 new messages