MultiPage input, single page output

[cflynt]

I'm using a slightly modified version of the Sample TEST_OCR/TEST_OCR.cpp file as a subroutine.

This is working OK on single-page files, but when I OCR a 4-page file the output is only one page.

Am I missing something in my calls?

int doOCR (char *input, wchar_t *languagePath, wchar_t *output) {
DWORD dwMaxLevel;
HRESULT hr;

// Set options for OCR
PXO_Options Options;

// New PXO_inputFields
PXO_InputFields InFields;
BSTR textout;

PXODocument Doc;
// Initialize the document.
// Replace NULL with your developer code and license key

OCR_Init(&Doc, 'SECRET', 'SECRET');

// Set the callback function
OCR_SetCallback(Doc,SampleCallback,(LPARAM)&dwMaxLevel);

// Load the specified PDF into the PXODocument input layer
hr = OCR_LoadA(Doc, input);

if (hr != 0) {
printf("Failed to load file. %s\n", input);
return -1;
}

Options.blacklist = NULL;
Options.whitelist = NULL;
Options.ImageFlags = OCR_Image_Autorotate; // check for minor rotation of images and correct it
Options.lang = PXO_English;
Options.raster_dpi = 300;
Options.RegionMode = OCR_Auto;
Options.DataPath = languagePath;
Options.accMode = 0; // this field is reserved and unused for now

hr = OCR_MakeSearchable(Doc, &Options, NULL);

SampleCallback(98,99,hr);

if ((hr != 0) && (hr != 0xc20a01f4)) {
printf("Make searchable failed. Error code: %xx\n", hr);
printf("TOERR: %xx\n", DS_x_TO_ERROR(hr));
}
else
{
// Save output layer to a PDF file
OCR_SaveW(Doc, output);
}
return (hr);
}

[cflynt]

Being a bit less incoherent.
When the input file to the previous subroutine is a document with 4 pages (my only multi-page test so far), the
resulting output file is converted to text, but the new pdf document contains only the first page.

[Walter-Tracker Supp]

I'm assuming the OCR_MakeSearchable() call returns no error?

Everything looks okay visually although I'll run a test with your code to make sure I'm not missing anything; in the meantime, maybe you could provide the PDF file, either as an attachment or by email to [email protected]? Does it do this with other PDFs or just the one?

[Walter-Tracker Supp]

Just a quick update; I tested your code with a 7 page document and had no problem with the output. I'll await the other information I asked in my previous post. You may also want to check the return value of OCR_SaveW, i.e.:

Code: Select all

pseudocode: 

hr = OCR_SaveW(...)
if (( hr != 0) ...)
{
   printf("Save failed, error %xx\n", hr);
}

-Walter

[cflynt]

The sample PDF file should be in your mailbox.
Thanks.

[Walter-Tracker Supp]

Got it and responded!

I could not reproduce the issue with the included file. I was able to OCR all pages. Have you reproduced it with other files? What is the return code (HRESULT) from OCR_SaveW()?

Also note that you were using a text-based PDF for OCR. This works but will result in loss of information (Text -> Rasterized Image of the Page -> OCR Text). If you were just using it as a quick test of OCR that's fine (and you probably already know this), but for real work this isn't really the best workflow

-Walter