Welcome to the Monday session on medical image processing, the interventional part, the second

part, sorry.

And what we considered last week was a very intuitive concept and a very easy to understand

concept.

Let me briefly hook up to the problem that we have considered.

What we are working on this summer semester is we want to build interactive medical imaging

systems or medical imaging devices.

And that means the doctor is working with the patient and he wants to get as much information

as much as, I think I have to reset, sorry.

So the doctor is working with the patient, he is using images and these images should

provide a lot of information for working with the patient.

And that means a lot of interaction, that means that we have to compute some features

out of the images, that we have to combine images, that we have to build up a virtual

space where he can, for instance, simulate things and many aspects with respect to these

things are important for us.

And in winter we haven't seen that much on pre-processing in terms of finding features,

for instance.

And that's why the first part of the summer semester is dealing with the problem, how

can we find some markers, some objects, some particular objects in the image automatically.

And one problem that is considered first in this context is how can I find edges, for

instance, or corners or some kind of points that are significant for further processing.

For instance, if I have an image like this here, here is the skull.

And then this here.

Whatever that is.

And we want to find, for instance, here that there is a contour line.

So how can we detect this contour line automatically?

And one idea is to look into the representation of images in the computer and how are images

represented.

Basically, images are represented by grids or matrices.

So we have here a matrix and here we find numbers 7, 9, 5, 0, 0, 1, 1, for instance.

And these numbers represent intensity.

So if you look at an x-ray image, at a computer tomography image, you see a two-dimensional

matrix in the computer and each entry represents an integer, for instance, and the integer

is associated with an intensity value.

And now we want to find an edge.

That means we have to consider this matrix as a two-dimensional function and we want

to find those points in the function where we have a jump from intensity value 0 to intensity

value 255, for instance.

So we look for points in the image where we have a high change of intensity values in

the neighborhood.

So in terms of a one-dimensional function, what does that mean?

That basically means I have a function like this.

Dominic, which color?

Red.

Is there a reason?

You fell in love during the weekend?

You can talk to us.

It's confidential.

All is recorded.

Oh, interesting.