Yes, that's my point. You can't rely on the depth map in the image metadata to be the differentiator because it can easily be faked with depth estimation.