Text Recognition

The Text Recognition feature is exposed via the CTextRecognition class.

Text Recognition assumes that the input image/page is already oriented correctly and that the recognition language(s) are already known/detected. The CTextRecognition class can be used with the following parameter classes:

• COcrPageParams: class that contains the parameters for full-page text recognition. It should be used when recognition is performed on full pages.
• COcrZonalParams: class that contains the parameters for zonal text recognition. This class allows to provide an array of zones with associated settings - COcrZonalSettingsArray.

For full-page text recognition, you can consult the code snippet below:

Full-page Text Recognition

CIDRS objIdrs = CIDRS::Create();

// Load the source image
CImageIO objImageIO = CImageIO::Create(objIdrs);
CImage objImage = objImageIO.LoadImage("path/to/image");

// set the ocr params to recognize japanese text using thread count set to half the cores available on the system.
COcrContext objContext = COcrContext::Create(Language::Japanese);
COcrPageParams objOcrPageParams = COcrPageParams::Create(objContext);
objOcrPageParams.SetThreadingMode(ThreadingMode::Limited);

// run text recognition
CTextRecognition objTextRecognition = CTextRecognition::Create(objIdrs, objOcrPageParams);
CPageContent objPageContent = objTextRecognition.RecognizeText(objImage);

// objPageContent contains now the recognized text

Zonal Text Recognition

For zonal Text Recognition, have a look on the code snippet below:

CIDRS objIdrs = CIDRS::Create();

// Load the source image
CImageIO objImageIO = CImageIO::Create(objIdrs);
CImage objImage = objImageIO.LoadImage("path/to/image");

// zone set1: set the ocr params to recognize japanese text in one area and japanese table in another area
COcrContext objContext1 = COcrContext::Create(Language::Japanese);
COcrZonalSettings objZonalSettings1 = COcrZonalSettings::Create();
COcrZoneDescriptionArray xZones1 = COcrZoneDescriptionArray::Create();

IDRS_RECT rcZoneA = { 0, 0, 100, 100 };
IDRS_RECT rcZoneB = { 100, 100, 300, 300 };
COcrZoneDescription objZone1Description = COcrZoneDescription::Create(ZoneType::Text, rcZoneA);
COcrZoneDescription objZone2Description = COcrZoneDescription::Create(ZoneType::Table, rcZoneB);
xZones1.AddTail(objZone1Description);
xZones1.AddTail(objZone2Description);

objZonalSettings1.SetZones(xZones1);
objZonalSettings1.SetContext(objContext1);

// zone set2: set the ocr params to recognize English text in one area
COcrZonalSettings objZonalSettings2 = COcrZonalSettings::Create();
COcrContext objContext2 = COcrContext::Create(Language::English);
COcrZoneDescriptionArray xZones2 = COcrZoneDescriptionArray::Create();

IDRS_RECT rcZoneC = { 300, 300, 1000, 1000 };
COcrZoneDescription objZone3Description = COcrZoneDescription::Create(ZoneType::Text, rcZoneC);
xZones2.AddTail(objZone3Description);

objZonalSettings2.SetZones(xZones2);
objZonalSettings2.SetContext(objContext2);

COcrZonalSettingsArray xZonalSettings = COcrZonalSettingsArray::Create();
xZonalSettings.AddTail(objZonalSettings1);
xZonalSettings.AddTail(objZonalSettings2);
COcrZonalParams objOcrZonalParams = COcrZonalParams::Create();
objOcrZonalParams.SetZonalSettings(xZonalSettings);

// run text recognition
CTextRecognition objTextRecognition = CTextRecognition::Create(objIdrs, objOcrZonalParams);
CPageContent objPageContent = objTextRecognition.RecognizeText(objImage);

// objPageContent contains now the recognized text