Consider the setup as shown above. When the dot that is masked is seen, the eyes overlap
the images such that the object occluding it overlaps with the dot. This is in the real 3D setup. When the same thing is tried
out using two 2D images our eyes fail to do so. Try it yourself!
