Why the image is being formed at the surface? Last week I encountered a very cool optical phenomenon. A photo was sandwiched between the two glass plate to make the paperweight. The cool part is the picture seems to have been pasted on the top surface (see the pictures). 
I am unable to guess why there is an real-image being formed at the top surface. And I say that real-image is being formed at the surface not a virtual one because for all purposes it seems that picture is placed at the surface doesn't matter how you turn it. 
The surfaces of the mirrors are flat except at the edges where they beveled (I suppose) to give the comfortable grasp.
 A: I could not address all issues with wltrup's answer in one comment.
Firstly, the reflection number 1 is not responsible for the the top image of the photo.
This is actually reflection of the environment, and would be the same independently of the picture below. Here, the light follows the path:
ambient->surface->eye
To actually capture an image of the photo, the light must hit the picture and go back to the eye.
The path the light takes to hit the photo doesn't matter, but once it hits some point p, rays emanate from this point to the eye. Because of the way refraction works, outgoing rays get more inclined with respect to the normal, and the emmiting point seems closer to the surface, an effect that is even more intense at angles close to the plane of the surface. So the top image is simply distorted in position because of refraction at sharp angles. You can search for the exact formula (apparent depth refraction) if you want.
The middle image is only seen through the sides, and in this case the distortion contracts the image laterally, not upwards (remember it makes the image closer to the surface)
For the bottom image, since one is looking from above, the lower image can't come to the observer purely from refraction, because all the light rays are directed downwards. This means the light reflects off the bottom of the glass. In his solution, wltup assumed the photo was semi-transparent so light could get through. This can seen not to be true. Consider that the bottom side of a transparent photo is its mirror image:
Now, what would happen is that the light rays emmanating from the downside of the photo(inverted) would reflect at the bottom before reaching(normal) at your eye.
But the image is actually inverted! (You can verify that the m is at the wrong side of the circle) So the downwards photo is normal, which means we have a double sided photo that looks right from both sides. To look different from each side it must be opaque. So what actually happens is that light comes from the environment from below, not above, to then hit the photo and create a image, which is reflected and distorted by refraction, so it finally hits your eye!
Wow, that was pretty long for such a simple question. Hope I didn't make it any more confusing then the necessary.
The effect not changing perceptibly with angle is likely due to the chosen refractive index for the material 
