Okay, so let's start. No overview today, we just continue in the text. Tomorrow I will

give a brief overview and bring you back to the storyline of the lecture. And hopefully

tomorrow we will end up with a few more students in the audience than this afternoon. Generally

tomorrow afternoon there is this student general meeting in the Audimax. So we try to cancel

all our lectures as much as possible. There is just one exception, the medical engineering

lecture I have to give because I invited a guest. And on Tuesday, no last Thursday it

was announced that there should be no lectures. So it's a little too late for us. But Ellie's

lecture tomorrow will be cancelled. Okay, just for your information. Good. So we finished

last week the chapter on x-ray imaging using flat panel technology and we have discussed

a few smart ideas to do defect pixel interpolation. And now we continue in the text and we look

into the MR imaging modality and we will discuss one problem that is still an important problem

in magnetic resonance imaging and that's how to deal with intensity inhomogeneities. Whatever

that is we will see a few examples, many examples and we will also see a lot of smart methods

to deal with image inhomogeneities in MR imaging. So what we are going to do, that's the program

for this and for the next week, we will a little bit look into the MR acquisition devices

and how they work. But I have to admit that I'm not a physicist or I have to say I'm not

a physicist and I have to admit that I don't understand all the methods that are applied

there so I give you a very rough overview and you should be happy with that and keeping

in mind that the lecturer has no idea is always a good indicator for the prior probability

that questions into this direction will come up in the oral exam. That means the probability

is rather high because I try to learn. Now you are totally confused, right? That's the

purpose. Then we will talk about the problem of bias and gain fields in MRI. These are

the problems that show up when you acquire images and these are the artifacts that we

have to deal with. Then I basically will explain to you that we again have to solve a problem

that from a mathematical point of view looks rather difficult. Last week we learned how

to divide by zero. You remember that? This time I will tell you or I will show to you

a solution to the problem given a sum of two numbers decomposed into the components uniquely.

So you observe a sum of two integers and we decompose the sum of two integers into its

components. Okay? You might say there are many, many solutions, yes, but there is one

solution that we are interested in. I will show you how to compute that. We will talk

about the mathematical modeling that is lying below this. We will talk about then a bunch

of methods and I will just list all these methods and it will be overwhelming for you.

It will just kill you all the different methods and ideas that we discuss here. But these

are very, very important methods and it is good to know these things. Good. So far the

program. So we will talk about magnetic resonance imaging and how are these images acquired?

These images are acquired by systems that look as nice as these ones. You have magnets,

you use a certain principle borrowed from physics to visualize things. And we have to

say that these systems are extremely expensive. So they are 2 million and more euros. They

look from a distance quite similar to CT devices. So sometimes people are not aware of whether

they got a CT or an MR image. So you always have to ask them was it noisy or not. If it

is noisy then they have been in an MR device because the magnets cause noise here. And

this shows the most recent three Tesla systems. The most recent has to be changed because

it is no longer the most recent because the slide is already 3 years old and today we

have already 9 Tesla systems, 7.5 to 9 Tesla systems. So this is the Magnetom, Vario, Trio

scanner. The family of MR scanners in Siemens is called Magnetom. It is a marketing name

so that is their contribution. And there is a lot of software in it and we will learn

what type of software is in MR scanners. And I told you maybe if you browse the web and

see the structure of our lab, I mean we have a bunch of PhD students focusing on MR imaging

and related problems to MR images, what you can do with it and how to improve MR images.

And of course we use the local environment to do top research so we have a lot of collaboration