Vous êtes sur la page 1sur 5

So the last thing that I want to

talk about in this module is a,


a tentative explanation of why we don't
see the world the way it really is.
And the answer to that question,
is what's called the inverse problem.
And I want to spend a few
minutes talking to you about it,
because in some ways this problem
really forms the theme of the course.
We're going to come back to it again and
again.
It is arguably from my
point of view at least,
the major problem that vision
faces in its evolution and let
me explain to you what the problem is and
then come back to reemphasize that point.
So, here's a diagram of the inverse
problem as it applies to luminance,
again remember luminance is a physical
measure of the intensity of
light that falls on the detector or
the retina and
the elicited perception that
follows that is called whiteness.
So, this is a diagram of
the world we live in.
It's obviously quite simplified,
but I think you'll get the point.
And that world comprises,
with respect to vision,
a source of illumination,
objects in the world, and atmosphere.
And all of those things combined to
effect the luminance that
actually falls on the retina.
So let me explain that in
a little bit more detail.
So, photons coming from the,
the, the sun or, or
some other light source of course,
could be a, a new source of light or
whatever, that the illumination
that's flowing on objects is
one thing you need to know about
to behave correctly in the world.
We need to know what the illumination is,
unless we be confused by when seeing
objects in different illuminations as
fundamentally different sorts of things.
That illumination falls on
the surfaces of objects.
Those surfaces all have
different atomic properties and
the result of those properties
is that some of the photons
are absorbed by the surfaces of objects,
and some are reflected.
And the photons that are reflected,

that are going to reach the eye


have to pass through an atmosphere.
You know we don't live in a vacuum.
And that atmosphere is going to alter,
by diminishing or
altering the wavelengths that
are transmitted to the eye.
That's called transmittance.
And the result of that, is that what
falls on the retina is a combination
of illumination, surface reflectance,
reflectance efficiency, function of
different objects, the physical
characteristics of the atmosphere on
a particular day, in a particular place,
at a particular level above the sea.
And all of those factors, which are the
factors that we need to know about,
to navigate and behave properly in
the world, are, in a very real sense,
hidden, from the visual brain,
because they're all conflated at
the level of the retinal stimulus.
So, what do I mean by that?
Well let's look at this little
diagram here at the bottom.
The combination of illumination,
reflectance, and transmittance.
And of course, these are only some
of the factors that are involved in
effecting the amount of light
that actually reaches your eye.
Those are the physical parameters that
we need in some sense to understand.
But they are all entangled
in the retinal stimulus.
That is, there's no way, logically,
to separate the contribution of
illumination, reflectance and
transmittance from the stimulus that falls
on the retina at the back of the eye.
And that entanglement is referred
to as the inverse problem.
That is, how do you get back from the
entangled stimulus to the contributions
of the different factors that reflect or,
or express the parameters of the world,
in a way that allows you to behave in
the world as if you understood them?
And understanding them by any logical
analysis of the retinal stimulus, or
any reverse engineering
of the retina stimulus,
is really not a, a logical possibility.
This inverse problem is, of course a, a
general term that applies to any problem,
a mathematical problem,
where you don't know
the factors in the equation in,
in sufficient, a, amount to know.

So if you think of
the equation Z equals X Y and
you only know one of the parameters,
and you want to solve for, for Z.
Well you, you just can't do it
if you only know one parameter.
That's obvious.
And I think it's just a more formal
way of expressing the inverse problem,
as it applies to vision.
So as I say, this is really fundamental,
but the same problem exists.
Not just for movements,
which is what we're talking about in
this diagram, but as well for geometry.
So for example, diagram of the eye,
stimulus falling on the retina.
And now consider the fact
that the three objects here,
that again represent just a, a, a, minor
sampling of the infinity of possibilities
that could be casting the same
stimulus under the retina.
These three possibilities
are of different sizes,
different orientations,
different distances from the observer, but
they're all leading to the same
projection on the retina.
So how is the visual agent,
us mainly, to know whether
the object that's making this stimulus
on the retina, is a nearby object,
a more distant one an object
that's oriented in one way or
another way,
an object that's small or large.
Again, all of these things are entangled,
inflated,
if you want to use that word at
the level of the retinal image.
And there's really no way to
disentangle them in a manner that
is either obvious or
even logically possible.
I I, made the point earlier about
motion is equally strangely seen.
This is another major aspect of
the visual qualities that we're aware of.
Motion consists of the direction, and
speed of objects moving
in the environment.
And in this example,
what we're going to see,
[COUGH] I'll throw it
into action in a second,
is that these three different sources,
physical sources of the environment.
Again, this diagram just is
a representation of the retina or

any other detecting surface.


The image that's cast on this on this
surface, on the retina, is exactly
the same arising from objects of different
sizes, different orientations, different
distances that are moving at different
speeds and in different directions.
So let me run that, and
you'll see what I mean after it's obvious.
So as the objects move in space, again, in
different directions at different speeds
and different orientations, they're
all casting the same object, the same
image is cast by all of these
objects onto the, onto the retina.
Let me show you that once more.
Again, same image on the retina, or really
the same sequence of images on the retina.
The sequence of images is the where
withal of the generation of our
perception of motion.
They are, are, they,
these different objects in space
moving at different speeds,
and in different directions.
Are, are casting the same
sequence of images on the web.
So, these are really fundamental problems.
They exist, not just for luminance, but
for color, for geometry, for motion.
And as we'll talk about in
the later module, or for
judging depth, distance and size.
Whatever the basic visual quality is that
we're considering, this ingress problem,
this entanglement of the parameters that
measure the physical characteristics of
the world are not separable in
the retinal, in retinal images
that would allow us to get
back [COUGH] from the images.
That form the retina stimuli to
the sources in the world that
we need to know about in some sense,
know in quotation marks, and
to, enables to react accordingly
with successful behavior.
So images are obviously giving us
something that allows us to succeed in
the world, but what that something is,
is not just the image per se,
because this conflation,
this entanglement really prevents
any understanding of the parameters
of the world, in a direct or even
an indirect way from the images, as such.
So let me just summarize the main
points that I've tried to get
across in this first session.
I think you will have been convinced by

the demonstrations that I showed you,


that the significance of images for
behavior in the physical world
is just inherently uncertain.
And as I say,
that's really thematically, the,
the challenge that we're
addressing in this course.
How is this problem dealt
with by animal vision and
by human vision in, in particular,
and how do our percepts in
some way inform us about how to behave in
the world given this inherent uncertainty?
So my second point is that the real
world is in that sense unknowable
by the direct logical operation of retinal
images as I've said a few times now.
That's just a basic conundrum.
And the question that's going to be
the one that we come back to again and
again in different modules, in different
ways in this course, is the question that,
how does vision succeed in a world that
is fundamentally hidden in a sense of
the unavailability of physical parameters
that we intuitively would think
are generating our subjective perceptions
of the world and how to behave in it?
So, that's all for today.
Next time, we're going to talk about
the organization of the visual system and
the nature of visual stimuli.
I, I, I, certainly we want to focus
on the instruction inflection of
the individual system, but obviously
we need to know something about it,
and I'm going to tell you what I think we
need to know and we need to understand
what the nature of visual stimuli
really are in in physical terms.

Vous aimerez peut-être aussi