When our vision centres have analysed the incoming signals from the retina and processed it into information about edges, lines, colours, objects - does that then get reassembled somewhere to make the image we see in our heads? Its a very intuitive idea but that can't be possible. Because if your brain is recreating the 'image' inside your head somewhere, something then needs to look at it (and process it and break it down again to information about edges, lines, colours). But you've already done that! You already have the data, no need to make it back into an image and then re-analyse it again (see Dennett, cartesian theatre, inner observer, homunculus etc).

This is a simple concept but its very hard to get a grip on it, it is so against our intuitive feeling for how these things should work.

I have an analogy with computer games - imagine rendering a frame of a 3D game, Doom or something. The game has information about the 3D structure of the room, where the walls are, where the entities are. Then it goes to the rendering pipeline and that data is used to draw on a 2D screen, creating a 2d image. The light from that image then hits your retina and your vision processing centres break the image down into data again - the 3d structure of the room as far as you can see it, where the entities are, colours. That data is then in your head (obviously not exactly the same data that the game had, but some analogous data structure in whatever weird internal format the brain uses). And thats it! That data is you seeing it! (But no, you think, surely that data is used to recreate the image in my head somewhere....)

Things like the necker cube gives us a peep behind the curtain maybe.

I saw something similar at sunset the other day, a luminous dog! A fluffy white dog: the last sunbeams hit him from behind, so that he had an amber corona.