The following content has been provided by the University of Erlangen-Nürnberg.

So, good morning everybody to the Tuesday morning lecture on diagnostic medical image

processing. Now we will go into more practical problems. So far we have discussed different

modalities, we had the general introduction, we had a very tough refresher course on SVD,

which obviously reduced the number of people in the audience and now we will more or less

look into very specific problems. So don't worry we will not have a course on applied

mathematics here, that was just an introduction yesterday and last week and now we will use

these methods and all the mathematics we learned in the first two years of our studies at the

university to solve practical problems. And for me it's very important to let you know

that you should not expect that those things that we discuss here will show up one to one

later on in your industrial career. What I really intend to teach here is the way of

engineering which is expected when you work for industry. So you get some challenging

problem, there is no obvious solution, you will not find a solution in the slides or

in any books and you are requested to solve the problem. And for doing so you are required

to be equipped with tough engineering skills, how to deal with a problem, how to simplify

a problem and how to come up with a practical solution. That's the intention that I follow

while teaching this course. I say it in an easier way, this way do not expect to get

200,000 euros a year by solving problems that you can look up in my manuscript. That

will not happen for sure. Either you end up with earning just let's say 100,000 euros

and you, well let's say 30,000 euros and you find all the problems in the script or you

really have to develop engineering skills to do things that others cannot do in this

simple way. And it's also important, you know, when you will be on the market as an engineer

you must be equipped with a skill set that is not so easy to copy, because otherwise

others will do the job for half the money. So you really have to learn and to build up

a skill set that you are able to do things that others cannot do the same way in such

a short time. So what are we going to discuss in the following? The winter semester course

is basically divided up into four topics as I mentioned already. We talked a little bit

about modalities and now we will talk about pre-processing. And the pre-processing routines

we will consider will basically deal with the problem, what can I do to the image to

make it look better. I acquire an image by using some physics, some detector technology

and so on and before I bring the image on the monitor for diagnostic purposes I do some

pre-processing. And we will learn specific problems that show up during the acquisition

procedure and we will learn how to eliminate the artifacts that are implied by that. And

I will explain to you what actually pre-processing and post-processing mean to us and how we

defer the two things. So we talk about image pre-processing and it's a very simple explanation

but it's very powerful. If somebody is talking about image pre-processing it just means between

the detector and the monitor there is some image processing going on. Post-processing

means you have acquired the image, you look at the image at the monitor and you say oh

it's still too noisy and then you apply some post-processing routines to make the image

look better. Or you try to find out a segmentation where you want to find edges or you want to

compute some areas or volumes or things like that. That's post-processing. So what we should

keep in mind whenever we talk about pre-processing that means on the way from the detector to

the monitor we do some pre-processing. And the pre-processing has to be done fast because

usually the doctor presses the foot switch, captures the image and then he wants to see

the picture in many cases. Okay? So there are obvious reasons why image pre-processing

is important. This is just for your information. I do not learn these things for the oral exams.

I cannot drink so much that I start to ask questions like that. Blah, blah, blah. But

it's nice to know a little bit about that. So the improvement of image quality has to

meet the expectations or requirements of the physicians. They want to see certain image

quality. If they capture images they don't want to see too much noise in the image. They