The LEADTOOLS OCR Library provides methods for incorporating optical character recognition (OCR) technology into an application. OCR is used to process bitmap document images into text.
After the LEADTOOLS .NET OCR toolkit is installed to the system, programming with LEADTOOLS OCR can begin. Note that OCR features must be unlocked before the OCR properties, methods, and events can be used. For more information about unlocking LEAD features, refer to Unlocking Special LEAD Features.
To start using LEADTOOLS for .NET OCR in your application, add references to the Leadtools.Ocr.dll and Leadtools.Document.Writer.dll assemblies in your .NET application. These assemblies contain the various interfaces, classes, structures, and delegates used to program with LEADTOOLS OCR.
The code that interfaces with the engine is stored in a separate assembly that will be loaded dynamically after an instance of the IOcrEngine interface is created. Hence, be sure the engine assembly used resides next to the Leadtools.Ocr.dll assembly. You can add the engine assembly as a reference to your project—if desired—to automatically detect dependencies, even though this is not required by LEADTOOLS.
LEADTOOLS provides methods to:
LEADTOOLS uses an OCR handle to interact with the OCR engine and the OCR document containing the list of pages. The OCR handle is a communication session between LEADTOOLS OCR and an OCR engine installed on the system. This OCR handle is an internal structure that contains all the necessary information for recognition, getting and setting information, and text verification.
Steps 4, 5, 6 and 7 can be done pretty much in any order, as long as the steps are carried out after starting up the OCR engine and before recognizing a page.
For more information on the engine assemblies, refer to OcrEngineType and Files To Be Included With Your Application .
OCR an image file (or a LEADTOOLS RasterImage object) and obtain the text with optional formatting and position information. In this mode, an IOcrDocument object is not needed since the result is not going to be saved. The IOcrEngine.CreatePage method can be used to quickly create an IOcrPage from the RasterImage directly, call the necessary method (such as IOcrPage.Recognize) and then obtain the text directly using IOcrPage.GetText or IOcrPage.GetRecognizedCharacters. For an example, refer to IOcrEngine.CreatePage.
Perform low-level Optical Character Recognition of one or more pages followed by creating a document in the final document format, such as PDF or DOCX. In this mode, the user generally creates an IOcrDocument object (in memory or file-based) and then adds IOcrPage objects to it. The pages can be recognized either beforehand or at some point afterwards. When all the pages are added and recognized, IOcrDocument.Save is called to convert the recognition data to the final document. For an example, refer to IOcrDocument.
Perform high-level Optical Character Recognition on an input image file and output directly to a final document format such as PDF or DOCX. In this mode, you can use IOcrAutoRecognizeManager to convert the document in one shot. Various events and logging mechanisms can be used to modify and track the recognition operation. For an example, refer to IOcrAutoRecognizeManager.
C#
// Assuming you added "using Leadtools.Codecs;", "using Leadtools.Ocr;" and "using Leadtools.Document.Writer;" at the beginning of this class
// *** Step 1: Specify the engine type and create an instance of the IOcrEngine interface.
// This example will use the LEADTOOLS OCR Module - LEAD Engine and use it in the same process
IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.LEAD);
// *** Step 2: Start up the engine.
// Use the default parameters
ocrEngine.Startup(null, null, null, @"C:\LEADTOOLS23\Bin\Common\OcrLEADRuntime");
// *** Step 3: Create an OCR document with one or more pages.
IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument();
// Add all the pages of a multipage TIF image to the document
ocrDocument.Pages.AddPages(@"C:\LEADTOOLS23\Resources\Images\Ocr.tif", 1, -1, null);
// *** Step 4: (Optional) Establish zones on the page(s), either manually or automatically
// Automatic zoning
ocrDocument.Pages.AutoZone(null);
// *** Step 5: (Optional) Set the active languages to be used by the OCR engine
// Enable both the English and German languages
ocrEngine.LanguageManager.EnableLanguages(new string[] { "en", "de" });
// *** Step 6: (Optional) Set the spell checking engine
// Enable the spell-checking system
ocrEngine.SpellCheckManager.SpellCheckEngine = OcrSpellCheckEngine.Native;
// *** Step 7: (Optional) Set any special recognition module options
// Set the zoning method for the first zone in the first page to be Graphic so it will not be recognized
OcrZone ocrZone = ocrDocument.Pages[0].Zones[0];
ocrZone.ZoneType = OcrZoneType.Graphic;
ocrDocument.Pages[0].Zones[0] = ocrZone;
// *** Step 8: Recognize
ocrDocument.Pages.Recognize(null);
// *** Step 9: Save recognition results
// Save the results to a PDF file
ocrDocument.Save(@"C:\LEADTOOLS23\Resources\Images\Document.pdf", DocumentFormat.Pdf, null);
ocrDocument.Dispose();
// *** Step 10: Shut down the OCR engine when finished
ocrEngine.Shutdown();
ocrEngine.Dispose();
The following examples demonstrate some of the different ways the LEADTOOLS OCR engine can be used.
OCR an image file (or LEADTOOLS RasterImage object) and obtain the text with optional formatting and position information. In this mode, an IOcrDocument object is not needed since the result is not going to be saved. Use the IOcrEngine.CreatePage method to quickly create an IOcrPage from the RasterImage directly: call the necessary method (such as IOcrPage.Recognize) and then obtain the text directly using IOcrPage.GetText or IOcrPage.GetRecognizedCharacters.
Note: This mode is supported only by the LEADTOOLS OCR Module - LEAD Engine. Calling IOcrEngine.CreatePage using any other OCR engine will result in an exception being thrown.
The following example uses an OCR page without a document.
C#
// Create the engine instance
using (IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.LEAD))
{
// Start up the engine
ocrEngine.Startup(null, null, null, @"C:\LEADTOOLS23\Bin\Common\OcrLEADRuntime");
// Load the first page as a RasterImage
RasterImage rasterImage = ocrEngine.RasterCodecsInstance.Load(@"C:\LEADTOOLS23\Resources\Images\Ocr.tif", 1);
// Create an OCR page from this image, transferring ownership of the RasterImage object
using (IOcrPage ocrPage = ocrEngine.CreatePage(rasterImage, OcrImageSharingMode.AutoDispose))
{
// Recognize the page
ocrPage.Recognize(null);
// Show the text of all zones
for (int zoneIndex = 0; zoneIndex < ocrPage.Zones.Count; zoneIndex++)
{
string text = ocrPage.GetText(zoneIndex);
Console.WriteLine(text);
}
}
// The engine will automatically shut down when Dispose is called
}
Saving OCR results to a final document such as PDF or DOCX requires an instance of IOcrDocument. One or more OCR pages can be added to the document and then the various Save methods can be called to create the final document.
IOcrDocument can be used in two ways:
In memory-based mode, the OCR pages are required to be in memory before saving. This is not recommended when the document has a large number of pages and either a file-based document or the LEADTOOLS Temporary file format DocumentFormat.Ltd is required.
In memory-based IOcrDocument, the IOcrPageCollection holds the pages. Any or all of the pages can be recognized at any time and pages can be added or removed at will.
The following example uses a memory-based document to create a multipage PDF file. Note how all the pages are kept in memory during the save operation.
C#
// Create the engine instance
using (IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.LEAD))
{
// Start up the engine
ocrEngine.Startup(null, null, null, @"C:\LEADTOOLS23\Bin\Common\OcrLEADRuntime");
// Create the OCR document in memory
using (IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument(null, OcrCreateDocumentOptions.InMemory))
{
string imageFile = @"C:\LEADTOOLS23\Resources\Images\Ocr.tif";
// Add all the pages to the document
ocrDocument.Pages.AddPages(imageFile, 1, -1, null);
// Recognize all the pages
ocrDocument.Pages.Recognize(null);
// Save the recognition results in PDF format
ocrDocument.Save(@"C:\LEADTOOLS23\Resources\Images\Document.pdf", DocumentFormat.Pdf, null);
}
}
In file-based document mode, OCR pages are not required to be in memory before saving. This mode is best when the document has a large number of pages.
In file-based IOcrDocument mode, the IOcrPageCollection stores only views of the pages. When a page is added, a snapshot of the current recognition data is saved into the document. This data cannot be modified any more and the page is no longer needed. Pages must be recognized before they are added to the document and pages can only be added. They cannot be removed.
The following example uses a file-based document to create a multipage PDF file. Notice that the pages are disposed after they are recognized and are not required during the save operation.
C#
// Create the engine instance
using (IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.LEAD))
{
// Start up the engine
ocrEngine.Startup(null, null, null, @"C:\LEADTOOLS23\Bin\Common\OcrLEADRuntime");
// Create a file-based OCR document
using (IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument(null, OcrCreateDocumentOptions.AutoDeleteFile))
{
string imageFile = @"C:\LEADTOOLS23\Resources\Images\Ocr.tif";
// Get the number of pages in the document
int pageCount = ocrEngine.RasterCodecsInstance.GetTotalPages(imageFile);
// Create a page
for (int page = 1; page <= pageCount; page++)
{
// Load a RasterImage
RasterImage rasterImage = ocrEngine.RasterCodecsInstance.Load(imageFile, page);
// Create an OCR page from this image, transferring ownership of the RasterImage object
using (IOcrPage ocrPage = ocrEngine.CreatePage(rasterImage, OcrImageSharingMode.AutoDispose))
{
// Recognize the page
ocrPage.Recognize(null);
// Add it to the document
ocrDocument.Pages.Add(ocrPage);
// Page will be disposed here and its memory freed
}
}
// Save the recognition results in PDF format
ocrDocument.Save(@"C:\LEADTOOLS23\Resources\Images\Document.pdf", DocumentFormat.Pdf, null);
}
}
File-based documents can also be saved and re-loaded to continue adding pages, or to convert it to the final document format at a later time. The following example shows how to do this.
C#
private static void Test4()
{
// Create the engine instance
IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.LEAD);
// Start up the engine
ocrEngine.Startup(null, null, null, @"C:\LEADTOOLS23\Bin\Common\OcrLEADRuntime");
string imageFile1 = @"C:\LEADTOOLS23\Resources\Images\Ocr1.tif";
string imageFile2 = @"C:\LEADTOOLS23\Resources\Images\Ocr2.tif";
// Create a file-based OCR document
// Pass a file name (the name will be re-used) and tell the document to not delete it
string documentFile = @"C:\LEADTOOLS23\Resources\Images\document.bin";
using (IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument(documentFile, OcrCreateDocumentOptions.None))
{
// Verify the document does not have any pages
System.Diagnostics.Debug.Assert(ocrDocument.Pages.Count == 0);
// Add a page
RasterImage rasterImage = ocrEngine.RasterCodecsInstance.Load(imageFile1, 1);
using (IOcrPage ocrPage = ocrEngine.CreatePage(rasterImage, OcrImageSharingMode.AutoDispose))
{
ocrPage.Recognize(null);
ocrDocument.Pages.Add(ocrPage);
}
// Here the document is disposed but the file will not be deleted
}
// Re-load the document
using (IOcrDocument ocrDocument = ocrEngine.DocumentManager.CreateDocument(documentFile, OcrCreateDocumentOptions.LoadExisting))
{
// Verify the document has one page
System.Diagnostics.Debug.Assert(ocrDocument.Pages.Count == 1);
// Add another page
RasterImage rasterImage = ocrEngine.RasterCodecsInstance.Load(imageFile2, 1);
using (IOcrPage ocrPage = ocrEngine.CreatePage(rasterImage, OcrImageSharingMode.AutoDispose))
{
ocrPage.Recognize(null);
ocrDocument.Pages.Add(ocrPage);
}
// Verify that the document has 2 pages
System.Diagnostics.Debug.Assert(ocrDocument.Pages.Count == 2);
// Save the document
ocrDocument.Save(@"C:\LEADTOOLS23\Resources\Images\Document.pdf", DocumentFormat.Pdf, null);
// The result will be a PDF file with two pages
}
// Finally, delete the document file
System.IO.File.Delete(documentFile);
ocrEngine.Dispose();
}
All of the previous techniques required low-level code to load a page, recognize it, and add it to a document. The LEADTOOLS OCR engines also support performing the same task above using the one shot "fire and forget" IOcrAutoRecognizeManager interface. In this high-level OCR, the input image is converted directly to the output format using the best options and just one method.
C#
// Create the engine instance
using(IOcrEngine ocrEngine = OcrEngineManager.CreateEngine(OcrEngineType.LEAD))
{
// Start up the engine
ocrEngine.Startup(null, null, null, @"C:\LEADTOOLS23\Bin\Common\OcrLEADRuntime");
// Convert the multipage TIF image to a PDF document
ocrEngine.AutoRecognizeManager.Run(
@"C:\LEADTOOLS23\Resources\Images\Ocr.tif",
@"C:\LEADTOOLS23\Resources\Images\Document.pdf",
DocumentFormat.Pdf,
null,
null);
}