I have a project where we are asked to assume the existence of a document D on a soft page which is handwritten in ink, a jpeg file of its native photograph (image A), and another discolored and angled/stretched image B in PDF format which shows the same document. We have to find if the image B was obtained by scanning the original document (D) or by printing and scanning its native jpeg image (image A).

The project directive does not want us to focus too much on stretching and the angle of image B. The primary focus is to be on discoloration (includes shadows as well which are heterogeneous on document B. The center of document B has a dark colored square and the discoloration in B is spread on a much larger area.

For the purpose of this project, we have been asked to assume that the printer and scanner are, in general, good quality machines, and that we are not allowed to see the metadata of any image. Is there any way to approach solving this?

  • 2
    I don't think that this is really an information security problem even though it is framed as forensics. The underlying problems are probably more appropriate for Data Science. Mar 20, 2021 at 10:02
  • This question has nothing to do with information security. I suggest to close it.
    – mentallurg
    Mar 20, 2021 at 12:39


Browse other questions tagged .