In article <757146996wnr at oimsdnp.demon.co.uk>, Peter at oimsdnp.demon.co.uk wrote:
So I stirred a hornets nest eh?
Not really, just trying to help you out. You have some misconceptions that
can be fixed pretty easily.
If I made any mistake, it was perhaps in misinterpreting the original
posters description of 'perception of 3D' with 'seeing' 3D.
Of course we can 'perceive' 3D by a number of means, but to 'see' it,
you need two eyes. If you don't believe me, look at a 3D object and
close one eye. You still know the object has depth because you know
from experience that it has, but you can't 'see' it.
The difference between perception and seeing ... let's leave that one be.
More concretely, I can take you to a room you've never been before, with
one eye patched. You will be able to navigate the room very successfully
based on (at least) the cues that I mentioned in my previous post, with no
binocular information. This does not rely on knowledge of the room. It
relies on cues like shading, occlusion, and movement that are physical
aspects of the scene. These can be used by a single camera connected to a
computer (for example), to get good estimates of the 3D structure of the
I recommend you to a large, empirical and sometimes thoughtful literature
on this topic.
ej at white.stanford.edu