I’m trying to understand what you are saying, but the reality is these documents are not printed to PDF as an image in the general case. They contain text with formatting which can be selected and copied from the PDF document. Is that not correct?
The image is formed by the PDF reader you choose, no?
Here, maybe this will help. The file that presents in the reader as an image is 1.4 MB. The file that presents as text is only 59 kB. So it would seem the text selectable PDF file is not an image at all. It contains text with font descriptions and location information.
I have seen PDF files where the software that generated it was not at all sophisticated and placed each and every letter as a separate entity. When the text is selected in the PDF viewer, the selection is not exactly contiguous, either selecting other text as if it were part of the string seen, or adding spaces between each letter, or both. This would seem to be further evidence that the PDF file is not an image file like a JPG or PNG, but a text file with formatting.