does not output correctly based on unicharset file.(Kannada)

17 views
Skip to first unread message

Sriranga(78yrsold)

unread,
Apr 11, 2011, 5:06:16 AM4/11/11
to indic-ocr, Debayan Banerjee
 version r-527 OS= win XP
Based on Tunga.txt generated TungaMap.tif and box file.
 Output i.e. test.txt  should  contains/display all chars shown in the  map.unicharset file.
But It is observed some of chars missing or misspelling. why it happens and what is the
solution?
Tunga.txt
TungaMap..tif
TungaMap.box
map.unicharset
map.traineddata
test.txt

Sriranga(78yrsold)

unread,
Apr 12, 2011, 12:07:50 PM4/12/11
to indic-ocr, Debayan Banerjee, Ray Smith
Dear Debayanin,
Here also same problem  with hindi(Devanagari) it is observed that contents of  output text viz. devtest.txt does not agree with contents(char/akshara) in dev.unicharset file - even though used same tif file used to generate traineddata as well as for ouputtest.
My contention is since the same tif was used for generating datafiles as well as for output testing and as such contents of the outputtest file should reflect  same chars/aksharas-without any changes- generated in the unicharset file.
The above same problem also existed in the Kannada -which already posted.
Awaiting valuable guidance/solution.
With Regards,
-sriranga(78yrs)



2011/4/11 Sriranga(78yrsold) <withbl...@gmail.com>
devanagari.txt
devanagari.tif
devanagari.box
devtest.txt
dev.unicharset
dev.traineddata
Reply all
Reply to author
Forward
0 new messages