Cleanup & reindent .cpp files

This commit is contained in:
Philippe Vaucher
2014-03-19 11:26:31 +01:00
parent e85651ef84
commit 460205e943
26 changed files with 2349 additions and 2805 deletions

View File

@@ -17,30 +17,26 @@
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "ocr.h"
OCR::OCR(Config* config)
{
this->config = config;
this->postProcessor = new PostProcess(config);
this->postProcessor = new PostProcess(config);
tesseract=new TessBaseAPI();
tesseract=new TessBaseAPI();
// Tesseract requires the prefix directory to be set as an env variable
vector<char> tessdataPrefix(config->getTessdataPrefix().size() + 1);
// Tesseract requires the prefix directory to be set as an env variable
vector<char> tessdataPrefix(config->getTessdataPrefix().size() + 1);
strcpy(tessdataPrefix.data(), config->getTessdataPrefix().c_str());
putenv(tessdataPrefix.data());
strcpy(tessdataPrefix.data(), config->getTessdataPrefix().c_str());
putenv(tessdataPrefix.data());
tesseract->Init("", config->ocrLanguage.c_str() );
tesseract->SetVariable("save_blob_choices", "T");
//tesseract->SetVariable("tessedit_char_whitelist", "ABCDEFGHIJKLMNPQRSTUVWXYZ1234567890");
tesseract->SetPageSegMode(PSM_SINGLE_CHAR);
tesseract->Init("", config->ocrLanguage.c_str() );
tesseract->SetVariable("save_blob_choices", "T");
//tesseract->SetVariable("tessedit_char_whitelist", "ABCDEFGHIJKLMNPQRSTUVWXYZ1234567890");
tesseract->SetPageSegMode(PSM_SINGLE_CHAR);
}
OCR::~OCR()
@@ -50,18 +46,14 @@ OCR::~OCR()
delete tesseract;
}
void OCR::performOCR(vector<Mat> thresholds, vector<Rect> charRegions)
{
timespec startTime;
getTime(&startTime);
postProcessor->clear();
for (int i = 0; i < thresholds.size(); i++)
{
@@ -69,7 +61,6 @@ void OCR::performOCR(vector<Mat> thresholds, vector<Rect> charRegions)
bitwise_not(thresholds[i], thresholds[i]);
tesseract->SetImage((uchar*) thresholds[i].data, thresholds[i].size().width, thresholds[i].size().height, thresholds[i].channels(), thresholds[i].step1());
for (int j = 0; j < charRegions.size(); j++)
{
Rect expandedRegion = expandRect( charRegions[j], 2, 2, thresholds[i].cols, thresholds[i].rows) ;
@@ -79,52 +70,54 @@ void OCR::performOCR(vector<Mat> thresholds, vector<Rect> charRegions)
tesseract::ResultIterator* ri = tesseract->GetIterator();
tesseract::PageIteratorLevel level = tesseract::RIL_SYMBOL;
do {
const char* symbol = ri->GetUTF8Text(level);
float conf = ri->Confidence(level);
do
{
const char* symbol = ri->GetUTF8Text(level);
float conf = ri->Confidence(level);
bool dontcare;
int fontindex = 0;
int pointsize = 0;
const char* fontName = ri->WordFontAttributes(&dontcare, &dontcare, &dontcare, &dontcare, &dontcare, &dontcare, &pointsize, &fontindex);
bool dontcare;
int fontindex = 0;
int pointsize = 0;
const char* fontName = ri->WordFontAttributes(&dontcare, &dontcare, &dontcare, &dontcare, &dontcare, &dontcare, &pointsize, &fontindex);
if(symbol != 0 && pointsize >= config->ocrMinFontSize) {
postProcessor->addLetter(*symbol, j, conf);
if (this->config->debugOcr)
printf("charpos%d: threshold %d: symbol %s, conf: %f font: %s (index %d) size %dpx", j, i, symbol, conf, fontName, fontindex, pointsize);
bool indent = false;
tesseract::ChoiceIterator ci(*ri);
do {
const char* choice = ci.GetUTF8Text();
postProcessor->addLetter(*choice, j, ci.Confidence());
//letterScores.addScore(*choice, j, ci.Confidence() - MIN_CONFIDENCE);
if (this->config->debugOcr)
{
if (indent) printf("\t\t ");
printf("\t- ");
printf("%s conf: %f\n", choice, ci.Confidence());
}
indent = true;
} while(ci.Next());
}
if(symbol != 0 && pointsize >= config->ocrMinFontSize)
{
postProcessor->addLetter(*symbol, j, conf);
if (this->config->debugOcr)
printf("---------------------------------------------\n");
printf("charpos%d: threshold %d: symbol %s, conf: %f font: %s (index %d) size %dpx", j, i, symbol, conf, fontName, fontindex, pointsize);
delete[] symbol;
} while((ri->Next(level)));
bool indent = false;
tesseract::ChoiceIterator ci(*ri);
do
{
const char* choice = ci.GetUTF8Text();
postProcessor->addLetter(*choice, j, ci.Confidence());
//letterScores.addScore(*choice, j, ci.Confidence() - MIN_CONFIDENCE);
if (this->config->debugOcr)
{
if (indent) printf("\t\t ");
printf("\t- ");
printf("%s conf: %f\n", choice, ci.Confidence());
}
indent = true;
}
while(ci.Next());
}
if (this->config->debugOcr)
printf("---------------------------------------------\n");
delete[] symbol;
}
while((ri->Next(level)));
delete ri;
}
}
if (config->debugTiming)
@@ -134,7 +127,4 @@ void OCR::performOCR(vector<Mat> thresholds, vector<Rect> charRegions)
cout << "OCR Time: " << diffclock(startTime, endTime) << "ms." << endl;
}
}