Question about pdf OCR

DEVONthink Tips
1

I recently scanned some documents into DTPO (edition 2.9.10.) using a Fujitsu Scansnap S1500M scanner and have found that certain of the pdfs I scanned (although by no means all of them) are not capable of being converted in DTPO so that they have searchable text. Note that I did not engage the OCR option in the Scansnap Manager when I scanned the documents. The only item differentiating the problematic pdfs from the others that I was able to OCR in DTPO following scanning was that those incapable of being OCRed are labelled in the “Kind” column of list view as “pdf” (in lower case) rather than “PDF” (i.e. in upper case) or “PDF+Text” (which, of course, indicates an OCRed pdf). See the attached screenshot - circled in red.

Can anyone explain the difference between files labelled in the Kind column “pdf” compared to files labelled “PDF” and why the former are seemingly incapable of being OCRed by DTPO? I have found a way to OCR the problematic “pdf” files by using the PDFPenPro app to “force OCR” each of them. However, even doing this and making them text searchable still leaves them with the lower case “pdf” label in DTPO rather than “PDF+Text”.

Many thanks in advance for any thoughts you might have. Please let me know if you require additional information from me to consider this issue.
Screenshot - pdf label issue.tiff (460 KB)

2

Scanned documents are images – the scanner itself knows nothing about text. OCR is the process of intelligently looking at an image and deciding what part of it is text, and what the text content is. Any OCR process adds a “text layer” to the PDF that is aligned with the text. In DEVONthink, Kind == PDF is a basic PDF image. No text layer. Kind == PDF + Text has, you guessed it, the image and the text layer.

If you have a PDF that isn’t getting OCRd when you ask DEVONthink to do so, it could be that the original scan is skewed badly or the image is bad quality so that the ABBYY engine just doesn’t understand the image. In any case, none of us can see (or want to see) “Jetts Haymarket”, so if would a good idea to open a ticket directly with Support (here or here or support_at_devontechnologies_dot_com), where you can get faster, more personalized attention than the forum. Attach a copy of the mis-behaving PDF to the ticket so they can examine it for you.

3

Many thanks korm. I will open a ticket.