Hello!

I am stucked in trying to figure out what is wrong with my code. I am doing 
exactly
the same as in comment in cube_object.h file:

// The typical usage pattern for Cube is shown below:////         // Create and 
initialize Tesseract object and get its//         // CubeRecoContext object 
(note that Tesseract object owns it,//         // so it will be freed when the 
Tesseract object is freed).//         tesseract::Tesseract *tess_obj =  new 
tesseract::Tesseract();//         tess_obj->init_tesseract(data_path, lang, 
tesseract::OEM_CUBE_ONLY);//         CubeRecoContext *cntxt = 
tess_obj->GetCubeRecoContext();//         CHECK(cntxt != NULL) << "Unable to 
create a Cube reco context";//         .//         .//         .//         // 
Do this to recognize a word in pix whose co-ordinates are//         // 
(left,top,width,height)//         tesseract::CubeObject *cube_obj;//         
cube_obj = new tesseract::CubeObject(cntxt, pix,//                              
                left, top, width, height);////         // Get back Cube's list 
of answers//         tesseract::WordAltList *alt_list = 
cube_obj->RecognizeWord();//         CHECK(alt_list != NULL && 
alt_list->AltCount() > 0);////         // Get the string and cost of every 
alternate//         for (int alt = 0; alt < alt_list->AltCount(); alt++) {//    
       // Return the result as a UTF-32 string//           string_32 res_str32 
= alt_list->Alt(alt);//           // Convert to UTF8 if need-be//           
string res_str;//           CubeUtils::UTF32ToUTF8(res_str32.c_str(), 
&res_str);//           // Get the string cost. This should get bigger as you go 
deeper//           // in the list//           int cost = 
alt_list->AltCost(alt);//         }////         // Call this once you are done 
recognizing this word//         delete cube_obj;////         // Call this once 
you are done recognizing all words with//         // for the current language// 
        delete tess_obj;//// Note that if the language supports "Italics" (see 
the CubeRecoContext), the// RecognizeWord function attempts to de-slant the 
word.
But I am obtaining wrong recognition results. Contrary, if I recognize the same 
image through TessBaseAPI interface with OEM_CUBE_ONLY turned on I am getting 
nice and correct results.
Should example from comment in cube_object.h work correctly? 

-- 
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