In my project,we need use OCR ,so I download some traineddata files from googole'code.But the result which they callback for me almost wrong.So I think I need do my traineddata.Why this traineddata which was created by myself does not work on android?Thx your help.
Add:
Thx to Richard MacCutchan'help first.Let me show more details about my question.
step one,I debug my app on OCR'project.Do it like this:
TessBaseAPI baseApi = new TessBaseAPI();
baseApi.init(getSDPath(), language)
Here, the field of 'language' is 'chi_sim',and I'm sure that I replace the file of chi_sim.traineddata to my
chi_sim.traineddata.However the result is so bad -my app crash.
My chi_sim.traineddata was produced step by step like this :
1.Create chi_sim.font.exp0.tif file.
2.Execute tesseract.exe chi_sim.font.exp0.tif chi_sim.font.exp0 batch.nochop makebox ,then get chi_sim.font.exp0.box file.
3.create font_properties file.include :
font 0 0 0 0 0
4.create a bat about :
rem font_properties
echo Run Tesseract for Training..
tesseract.exe chi_sim.font.exp0.tif chi_sim.font.exp0 nobatch box.train
echo Compute the Character Set..
unicharset_extractor.exe chi_sim.font.exp0.box
mftraining -F font_properties -U unicharset -O chi_sim.unicharset chi_sim.font.exp0.tr
echo Clustering..
cntraining.exe chi_sim.font.exp0.tr
echo Rename Files..
rename normproto chi_sim.normproto
rename inttemp chi_sim.inttemp
rename pffmtable chi_sim.pffmtable
rename shapetable chi_sim.shapetable
echo Create Tessdata..
combine_tessdata.exe chi_sim.
pause
5.execute this bat step 4,and I'm sure there no any mistake where this bat execute .
In the end,my engilish is not good . If grammar was wrong ,please point it.Thx your help.