Hi Raj, 

   ¿is there a possibility you could send me your language "sun" to test 
the Seven Segment OCR out?

I am unsure at the moment which is the best way to attack this problem, 
either with Tesseract or OpenCV and the fact that your solution is giving 
100% accuracy on your tests gives me hope for Tesseract.

Kind regards, 

   David 

El jueves, 9 de abril de 2009 08:14:04 UTC+2, Raj escribió:
>
> HI Paul, 
>
>  Now, i have got tesseract to working to recognize 7-segment digits. 
>
> and it is working perfectly as required producing 100% accurate 
> results with the command "tesseract.exe input.tif output.txt -l sun" 
> where "sun" is the language name. 
>
> But my intension is to use the tesseract in a c# application.for this 
> I found, EmguCV 1.5.0.1wrapper for OpenCV, an Example in C#  where the 
> example named(OCR) is calling the tesseract internally using tessdata 
> for English language,and Tessnet2 dll (a .NET 2.0 Open Source OCR 
> assembly using Tesseract engine)and it is working perfectly. when i 
> provide the image of 7-segment digits it is working with 90% accuracy 
> with some ambiguity. 
>
> For this i create my own 8 files of tessdata and tried to use it in 
> the example.But when i replace the tessdata in the example with my 
> tessdata files, the application is terminating at the statement 
> "ocr.init()" without producing any results. 
>
> But the application with tessdata with english language is working 
> perfectly as i said. 
>
> I have no clue of this prob. B'coz when i use my tessdata and run the 
> command "tesseract.exe input.tif output.txt -l sun" it is producing 
> results with cent% accuracy. But when i tried to use my tessdata in 
> the application, it is not producing any results. 
>
> I hope i'm clear to u abt the prob. 
>
> Please provide some feedback. 
>
> Thanks in advance, 
>
> Regards.. 
> Sunder Raj.M 
>
>
>
> On Apr 1, 9:52 pm, paulfeakins <[email protected]> wrote: 
> > Hi there, 
> > 
> > Have you managed to get tesseract working with 7-segment digits yet? 
> > 
> > I'd recommend downloading a "digital" font that looks right and 
> > training it by following this: 
> > 
> > http://code.google.com/p/tesseract-ocr/wiki/TrainingTesseract 
> > 
> > Paul.

-- 
You received this message because you are subscribed to the Google
Groups "tesseract-ocr" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/tesseract-ocr?hl=en

Reply via email to