I could be wrong, but I'm going to go with "not possible."
What you need to do to pull this off is make all the objects in a frame the same scale, which is not how the camera captures it. So you would need to selectively enlarge or shrink objects based on their distance from the camera. And the degree of scaling would depend on the distance from the camera and that information is not recorded in the image (you can only solve for the distance of the focus point based on the focal length of the lens and the position of the focus ring).
You would need a lot of rulers/yardsticks/etc in the image to get the correct measurements or you could measure the items and jot them down in a book. While you're at it (if the image allows for it) you might want to shoot all the subjects separately so that you can get all the subjects in focus (if that's what you want) and so you can manipulate them individually before layering them on top of each other.
I'd love for someone to prove me wrong, but I don't think it can be done with today's cameras.
Edit: if you're starting with a normal photo you could estimate the sizes of things and selectively stretch in photoshop, but that'd take a lot of work and I don't think you'll end up with a good result. I think the biggest challenge with this is the ground as it all needs to be stretched at varying rates as you move away from the camera.