Service to detect if a PDF is scanned?

[Aftia_Jeff]

Hi,

Is there a service that will detect if a PDF is scanned?  I'd like to determine if a PDF is text-based or a scanned image before OCRing it.   I don't see what I'm looking for in PDFProperities.

Thanks,

Jeff

[Joel Geraci · Accepted solution]

In the output from PDF Properties API, look in the "pages" property. For each page you'll see something like the code below.

Be sure to verify the "is_scanned" boolean by checking if the file has only one image and "only_images" is true. If the file has been OCRed, "has_text" will be true. 

{
    "page_number": 0,
    "is_scanned": true,
    "width": 630,
    "has_structure": false,
    "content": {
        "number_of_images": 1,
        "only_images": true,
        "has_text": false,
        "has_images": true,
        "is_empty": false
    },
    "height": 810
}

[Aftia_Jeff]

Great thanks Joel.

[Aftia_Jeff]

@Joel Geraci  We are working on implementing code for this now.  As you pointed out, is_scanned is a property on a page.   We plan to check if the first page is scanned and then OCR the PDF if it is.   Does that make sense?  The OCR service only works on the document level, there are no page parameters.