In this course, you will learn the science behind how digital images and video are made, altered, stored, and used. We will look at the vast world of digital imaging, from how computers and digital cameras form images to how digital special effects are used in Hollywood movies to how the Mars Rover was able to send photographs across millions of miles of space. The course starts by looking at how the human visual system works and then teaches you about the engineering, mathematics, and computer science that makes digital images work. You will learn the basic algorithms used for adjusting images, explore JPEG and MPEG standards for encoding and compressing video images, and go on to learn about image segmentation, noise removal and filtering. Finally, we will end with image processing techniques used in medicine. This course consists of 7 basic modules and 2 bonus (non-graded) modules. There are optional MATLAB exercises; learners will have access to MATLAB Online for the course duration. Each module is independent, so you can follow your interests.
4 - Human visual system - Duration 17:10
Loading...
Del curso dictado por Duke University
Image and Video Processing: From Mars to Hollywood with a Stop at the Hospital
372 calificaciones
Duke University
372 calificaciones
De la lección
Introduction to image and video processing
Learn what is image and video processing. Learn the very basic concepts of human perception needed for understanding image processing. Learn simple tools in signal processing needed to understand following units.
Conoce a los instructores
Guillermo SapiroProfessor
Electrical and Computer Engineering
Without any doubt, one of the main clients of image and video process is us.
So let's just learn something about our own visual system.
Here, we have a very simplified cross section of the human eye.
We have the cornea, here, we have the lenses and we have the retina.
The retina is where the images that we see,
basically images that come to our eyes are projected on.
So basically, as we see in this figure, we have the scene.
So this is the actual scene in the real world and it's projected into our retina,
here in the back of our eye. The retina basically has sensors,
as we are going to talk in a second. And the images are projected onto the
retina, and then they are sent into the brain.
Okay? So the retina is full of sensors,
all over the retina. And naturally we have two type of
sensors. We have cones.
The cones are basically marked here by the continuous line,
that's the density of cones that we have in the retina.
And as we can see, there is a very, very high density of cones.
We see a very, very high peak of cones around the fovea.
The fovea is basically where we can see the best.
Cones are very good at seeing details. Cones are very, very good at seeing in
bright light. So, we are, without us noticing of
course, in our regular scenario, we're always trying to move our eyes,
such that basically the scene is projected as much as possible into the
fovea region of our retina. So those are the cones.
Okay? And they, actually, their density the amount of cones, the amount of
receptors goes down when we get away from the fovea.
There is a different type of receptors basically of sensors, which are called
the rods and those are, sorry? The rods are basically marked in this
picture with these lines. So the concentration of the rods is
actually a big more uniform across the whole retina.
Basically, as we can see in this picture, there is of course, a very low, basically
nonexistent concentration where the cones take over and the rods are actually very
good to, not to look at details, but to get a general idea of what's happening at
the scene and they're actually very good to see in very dim light.
Okay? So, between the, the cones that are very
good at seeing at bright light and very, very concentrated around the fovea and
the rods which are spread all around the retina and are very good at seeing at
very low light. We get an extremely large spectrum of
light that we can observe with our naked eye.
Before I detail a bit more about that, I want to tell you that there's a region
in the retina that has basically no sensors and that's that region here,
that's called the blind spot. Basically, there's no receptors there and
we're going to notice that. And we're going to discuss that in a few
weeks when we talk about an image processing technique calling painting,
which as we have seen in the first videos is about making, removing objects from
the scene. We're going to be talking about the blind
spot a bit more and why or how. It's believed that we actually don't
notice that there is a region of our vision system that basically there are,
there are no reception of what's happening in the scene, but we are
completely unaware of that. So, going back to the point that we have
an extremely, extremely broad range of intensities that we can see when we
combine the cones and the rods, we should note that as depicting this figure.
We are actually, we can actually perceive a very, very broad range of intensities.
So this is in logarithmic, okay? It's really a huge part of intensities
that we are able to perceive and some of it is actually perceived by the rods,
these low intensity. Some if it is actually perceived by the
cones when we start going into high intensities.
One of the issues is that we're able to perceive very, very large range of
intensities, but not at the same time. So we cannot see in very dark rooms and
very light regions of the room at the same time.
Okay? So we, we can see very dark,
we can see in very dark rooms, we can also see in very light rooms. We cannot
see at the same time and what we need is actually to adapt. So at certain level
of, at certain level of intensity, for example, this point, we are actually
adapt and we are able to see this range. If we are adopting around here, we are
not able to see or at least not see very easily, around here or very, very low
intensities. That's what happens. We experience this
all the time when we're going to a dark movie theater. At first, we don't see
very well and after a while we adapt, and we actually see where we are going, and
where do we need to see. And the same when we go into a light
room. Now,
it's very important that, regardless of the background light, we're able to
distinguish the details and, and at, at different levels.
So when we go into a movie theater and the lights are off,
we don't want to see a uniform movie theater.
We need to find where basically, our seats are.
So we need to be able to discriminate even between very low levels of light and
we also need to be able to discriminate between very high levels of light.
So, that's what this adaptation allow us to do.
And there are very, very good experiments in the, in the literature to try to
understand, what kind of discrimination we can obtain at different levels of
light. And these experiments are basically very
simple in its, in its most fundamental form.
So there's a background light, and we basically draw a circle in the middle,
and we start changing the light of that circle until the observer says oh, I
notice that there is a circle in there. We can go up, making it lighter.
We can go down, making it darker. And then we measure, how much change are
we able to perceive? Now, the amount of change that we need to
perceive, depends, actually on the background light,
and that's what's called Weber law and this is shown in this picture.
It's a very interesting effect that if we are in very low light conditions,
we need the delta i, the change to be relatively high. If we are in very high
background light, we don't need so much change.
So the basic idea is that if we are in a very dark room, very small changes in, in
light, very, very small changes in, in, in brightness will not be perceived.
We actually need much wider range of changes, basically a, a much larger
change. So, basically two things that are very,
that both of them are very dark and very similar, we won't be able to distinguish
between them. On the other hand, when the background is
already very bright, we don't need so much difference between the different
levels to be able to distinguish between different objects.
And that, as I say, is called the Weber's Law and it's extremely important in the
design of images, and in the design of cameras, and design of different image
acquisition devices. Because if we want to distinguish between
two objects that both of them are very dark, we need to be very careful.
Because we know that we need a very large difference between those objects.
We are actually going to learn in Week 3, actually, how to make these differences
even larger so we can distinguish them. So if we take a picture of two very dark
objects that are different objects, but are basically a brightness is very
similar, we're going to be able to process that in the computer, so we can,
even with our naked eye distinguish between them.
So, the conclusion of this part is that we are able to observe and understand and
make interpretation of images in a very, very wide range of brightness,
but we are not able to do all that at the same time.
We need to adapt, and if we are adapting to dark, then we are less sensitive that
if we adapted to bright areas, to, to, very high levels of, of light.
Another very interesting aspect of our visual system is that our visual
perception depends on what's surrounding us.
One example is already Weber's law, if we can perceive this circle in the
middle depends on actually the brightness, the luminance, around it.
The other is what's called Mach bands. And it might be a bit difficult to see
this through the video, but let me explain it and in the next slide, we're
going to see an even more clear example. The basic idea is that white are 2% here.
I present you here a black region, a bit lighter, a bit lighter, a bit
lighter, and so on. The actual intensities that I'm painting
each one of these is, is constant. So basically, the intensity here is
constant as we can see it here. The intensity here is constant, is this
intensity, and so on.
So the intensity of every single one of these segments is constant.
Our perception actually, doesn't look like this, it's actually, doesn't look
constant like in this. It's actually like here.
We perceive that when a dark region is closer here, so we are at the boundary
between here, we actually perceive that the dark is even darker, and the bright
is even brighter. You might be able to perceive that in
these regions. For example, that basically, we believe
that because we're coming here into brighter region, we're coming from this
side, it actually looks darker. We perceive it as being darker.
It's not actually, but we perceive it as being darker.
I'm going to erase my annotation so maybe you can see that a bit better.
Let me just point again to where you should be looking at.
You should be looking at this boundary, for example.
I'm going to draw on top of it and then erase it,
you should be looking at this boundary that, when you're coming from the left,
it looks actually darker than it actually is.
And when you're coming from the right, it looks actually brighter than it is.
So let me just erase this and give you couple of seconds to try to see if you
see that. But if you don't, don't worry.
We're going to go into to the next slide and it's going to be even more clear.
Here, let me explain what we have. We have three rectangles that actually
have exactly the same brightness. So, they were painted with exactly the
same paint. Now, the background have, were painted
with three different levels of gray. This is dark, this is light, and this is
in between. And I hope you can perceive, even through
the video, that. Let me just go back.
You got this, looks darker than this one, because this
one, I'm going to erase all my annotation in a
second, so you can look at that again. This one is surrounded by a brighter
background than this one that is actually surrounded by a very, very dark
background. And then the perception is that this one
is actually perceived as being darker than this one.
Let's see if we can see that once more. Okay?
I hope you can perceive that. So there's a lot of visual illusions that
and, and tricks that our visual system plays with us and here is yet one more.
The idea here, there are a number of things.
Once again, I'm going to use my annotation,
then I'm going to erase it. So, I hope that all of you can perceive
here a square. This actually not a square,
but you can perceive, you believe that there is a square.
So you believe that there is a square right there,
but you, actually, there is not. There is actually, absolutely no
difference between the colors when I go around here, between the level of y.
There is actually no difference here. But, this gives us the illusion that
there's actually a square here. I hope you can see it now.
Similarly, in this image, we actually have straight lines coming in.
Nothing in the middle, but when I'm going to erase that. I hope
you will be able to perceive that there is kind of a circle form in the middle.
So we perceive. It's actually very interesting to see
that sometimes the actual responses of our visual system.
So if it were a measure, the actual firing of the neurons in certain regions
of our visual system, they're identical to those that will
happen if I actually drew a circle there. So, this is a visual illusion that is as
strong as an actual drawing. here is yet another example,
that because here we actually make an angle like this, and here we make an
angle like this, most of us will perceive this line as being longer than this line.
The same on the figure on the right. Most of us will perceive that these lines
are actually not parallel, but they actually are and the perception
that they are actually meeting each other is a result of basically a visual
illusion. So it's important when we're doing image
processing to understand how our basic understanding of the visual system works.
Once again, because the main client of image and video processing is basically
us. So, in the next video, we're going to
start seeing formulas. So far, I have shown you basic images and
some basic understanding of the human visual system really basic understanding.
Some people spend their whole careers studying the visual system, so these are
just some very, very basic process that are important for us to continue.
And now we are going to start talking about what are pixels, and what's
condensation, and what's quantization, and what's discretization,
and what are those megapixel cameras or even gigapixel cameras.
And what does all that mean. Thank you.
Explora nuestro catálogo
Inscríbete de manera gratuita y obtén recomendaciones personalizadas, actualizaciones y ofertas.
Coursera brinda acceso universal a la mejor educación del mundo, al asociarse con las mejores universidades y organizaciones, para ofrecer cursos en línea.
© 2018 Coursera Inc. Todos los derechos reservados.
