public static byte[] OCRInner(byte[] byteImages, string strLanguageFileName, int nFileFormat)
{
Dynamsoft.DotNet.TWAIN.DynamicDotNetTwain twain = new Dynamsoft.DotNet.TWAIN.DynamicDotNetTwain("586300F2E000E88E9965483ECB6CCD4C");
twain.OCRTessDataPath = StrPath;
twain.OCRLanguage = strLanguageFileName;
twain.OCRDllPath = StrPath + "bin";
switch (nFileFormat)
{
case FILE_TYPE_TEXT:
case FILE_TYPE_DOCX:
twain.OCRResultFormat = Dynamsoft.DotNet.TWAIN.OCR.ResultFormat.Text;
break;
case FILE_TYPE_PDF:
twain.OCRResultFormat = Dynamsoft.DotNet.TWAIN.OCR.ResultFormat.PDFPlainText;
break;
}
twain.LoadImageFromBytes(byteImages, Dynamsoft.DotNet.TWAIN.Enums.DWTImageFileFormat.WEBTW_BMP);
Dynamsoft.DotNet.TWAIN.IndexList tmp = new Dynamsoft.DotNet.TWAIN.IndexList(0);
return(twain.OCR(tmp));
}
public static byte[] OCRInner(byte[] byteImages, string strOCRLanguage, int nFileFormat)
{
Dynamsoft.DotNet.TWAIN.DynamicDotNetTwain DNObject = new Dynamsoft.DotNet.TWAIN.DynamicDotNetTwain("/*no License available*/");
DNObject.OCRTessDataPath = StrPath + "Resources\\ocr";
DNObject.OCRLanguage = strOCRLanguage;
DNObject.OCRDllPath = StrPath + "Resources\\ocr";
if (nFileFormat == -1)
DNObject.OCRResultFormat = (Dynamsoft.DotNet.TWAIN.OCR.ResultFormat)0;
else
DNObject.OCRResultFormat = (Dynamsoft.DotNet.TWAIN.OCR.ResultFormat)nFileFormat;
/*
* The uploaded image stream is in PDF format, so we are loading it as a PDF
*/
DNObject.LoadImageFromBytes(byteImages, Dynamsoft.DotNet.TWAIN.Enums.DWTImageFileFormat.WEBTW_PDF);
/*Fix the resolution for OCRing*/
for (short index = 0; index < DNObject.HowManyImagesInBuffer; index++)
{
Bitmap img = (Bitmap)DNObject.GetImage(index);
// If the PDF page says 72 dpi, and assume it's A4(8.0 X 11.0), then the page size is 612 X 792
// This way, we can computer the resolution to get the correct value(s)
int iTempImageResolution = (int)((float)img.Width * 72) / 612;
int iImageResolutionInteger = (int)((iTempImageResolution / 100) * 100);
int iImageResolutionMod = iTempImageResolution - iImageResolutionInteger;
int iImageResolution = iImageResolutionInteger;
if (iImageResolutionMod >= 80)
iImageResolution += 100; // e.g. 299
else if (iImageResolutionMod < 20)
{
// discard, e.g. 301
}
else
iImageResolutionInteger += 50;
// if we are not sure whether the resolution is valid, we can do the following simple check
/*100, 200, 150, 300 and 600 are the normal resolutions used to scan, add more if you use different ones*/
bool bResolutionValid = ((img.HorizontalResolution == 100) || (img.HorizontalResolution == 200) ||
(img.HorizontalResolution == 150) || (img.HorizontalResolution == 300) || (img.HorizontalResolution == 600));
if (!bResolutionValid)
{
//the resolution is missing, so we re-apply it
img.SetResolution(iImageResolution, iImageResolution);
DNObject.SetImage(index, img);
}
}
//Now we do OCR on the corrected images.
Dynamsoft.DotNet.TWAIN.IndexList tmp = new Dynamsoft.DotNet.TWAIN.IndexList();
for (short i = 0; i < DNObject.HowManyImagesInBuffer; i++)
{
tmp.Insert(i, i);
}
return DNObject.OCR(tmp);
}
