Simple test about self training data, but the result is so strange.

55 views
Skip to first unread message

Curtis0624

unread,
Nov 21, 2016, 12:41:09 PM11/21/16
to tesseract-ocr

Hi,
Attached is the target image (Image61.tif, characters are"330") and all train data(.traineddata, .box and .tif).


Information:

1.Windows 7.

2.Tesseract-OCR 3.02 or 3.05.00dev .

3.Command line:

Tesseract Image61.tif result -l fontyp


Specification:

1.traindata1\fontyp.traineddata

Trained data just includes "3". The result is 333.
2.traindata2\fontyp.traineddata
Trained data includes "3" and "5".
The result is 355 and why the second character isn't "3"?

Could anybody help me? Thanks a lot!!


SimpleTest.zip

Curtis0624

unread,
Nov 21, 2016, 10:50:20 PM11/21/16
to tesseract-ocr
Help~~

Curtis0624於 2016年11月22日星期二 UTC+8上午1時41分09秒寫道:

Curtis0624

unread,
Nov 23, 2016, 3:57:53 AM11/23/16
to tesseract-ocr
Please help me!!

Curtis0624於 2016年11月22日星期二 UTC+8上午1時41分09秒寫道:
Reply all
Reply to author
Forward
0 new messages