Error while running tesseract for a new traindata
Closed this issue · 3 comments
GoogleCodeExporter commented
What steps will reproduce the problem?
1.Collecting all the files for traindata
2.Making traindata
3.Put the traindata in tesssdata folder and run tesseract.
What is the expected output? What do you see instead?
Expected ouput is a text file containing the images of the text.Instead, I see
the error
Index>=0 &index<size_used:Error:Assert_Failed
Please use labels and text to provide additional information.
I gave the screenshot of the error and other files.My language name is ban and
font name is sl.
Original issue reported on code.google.com by m.tawfi...@gmail.com on 29 Mar 2015 at 4:03
Attachments:
GoogleCodeExporter commented
I think there is a problem with your font_properties file. It seems to have a
blank line above, while blank line should be at the end.
I was able to generate the traineddata with your files in jtessboxeditor (I
needed to add the words list, frequent words list and rename the font
properties file to the naming convention needed by the program.
BTW, there is already traineddata for Bangla - please see
https://code.google.com/p/tesseract-ocr/source/browse/ben.traineddata?repo=tessd
ata
and also see
https://code.google.com/p/tesseract-ocr/source/browse?repo=langdata#git%2Fben
Original comment by shreeshrii on 30 Mar 2015 at 8:50
GoogleCodeExporter commented
No, this will not work if I do not leave a blank space in front of the first
line, however, I have the same tif file as input.By the way,
Original comment by m.tawfi...@gmail.com on 31 Mar 2015 at 2:27
GoogleCodeExporter commented
You did not follow instruction[1] e.g. font_properties.txt does not meet
"Requirements for text input files", so I guess you did not created valid
traineddata.
Anyway you issue is invalid, because for support you should use tesseract user
forum. Issues tracker should be only for reporting of google produced
traineddata files.
[1] https://code.google.com/p/tesseract-ocr/wiki/TrainingTesseract3
Original comment by zde...@gmail.com on 9 Apr 2015 at 8:06
- Changed state: Invalid