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

Originally I was not or I'm still not required to teach this course, but I try to teach it

as often as possible. So if I am not available, Eva Kollertz will take over the course and

teach. And we have basically within this course four larger chapters and the chapter on image

reconstruction will be taught by two former lab members that are now working for Siemens.

So I will basically not be in charge of the lecture, let's say from December 1st to mid

of January or something like that, just to warn you. So for those of you who don't know

me, I think most of you know me, right? My name is Joachim Honecker, I'm here in charge

of the wellness in the 9th floor and in my spare time I do medical image processing and

I like that very much and this semester I try to teach as much as possible on diagnostic

medical image processing. First of all the question is what does it mean diagnostic medical

image processing? Diagnostic medical image processing basically means, and I always think

in very simple terms, patient comes to the hospital, the doctor has no time, the technician

takes care of the patient, captures a few images, prepares the images for the doctor

and the next day the doctor looks at them and tries to find out what's the problem with

the patient. That's basically a very rough interpretation of diagnostic medical imaging

or medical image processing. The images are acquired without a time pressure, in general

without a time pressure, then there is some time to work on the images and to prepare

the images for the final decision that will be done by the radiologist for example. And

the consequences for the algorithms we are considering here are basically that we should

not discuss too much efficiency, workflow embedding and things like that because this

will be part of summer semester where we will consider interventional image processing.

So that's the rough idea and I have prepared a few slides here. Before we go into the motivation

of all this, a few questions that show up all the time. When I start the lecture, first

question, can we shift the time for the lecture because I have collisions? The answer is no.

Very simple. Because if I shift the lecture for five people that are not available, another

five won't be available for the new slot and it's a never ending story. And here we are,

welcome. What's your name? Felix. Felix, welcome. But the lectures are recorded, you can download

them from iTunes, you can reconsider all my jokes and all the errors I do and all the

stupid things I say and it's very interesting. Two weeks ago I had oral exams and one student

showed up and he told me, your face is so familiar, I saw you every day on the videos

and so on and my response was, I haven't seen you before. Because this guy was never in

the lecture. Who cares? I don't care. Basically ask my questions, if I get the right answer,

you will succeed with a grade that you deserve. Is diagnostic medical image processing required

to attend part two? That's a very serious question. Why? Because I think, if you're

highly motivated, you can deal with IMIP without diagnostic medical image processing. My experience

is even if you're highly motivated, you have always the feeling while attending IMIP that

you missed some important points. That's true, right? Somehow. Do I need to know MATLAB or

Octave? Octave is the free version of MATLAB. It's a little different but basically covers

most of MATLAB's functionality. It's an advantage. My medical engineering friends here, they

all know it. But we will provide a tutorial and you will get used to it by the time. And

I don't want to miss the opportunity to stress here that the exercises are very, very important.

It's my personal experience and I think my students share this with me. You really understand

algorithms once you have implemented these algorithms. You can talk a lot about algorithms

and how fantastic they are and what nice math is involved. But to understand the algorithm,

it's important to implement it. All of you should have implemented a Fourier transform

because without having implemented Fourier transform, you don't see the core problems

basically associated with it. My English is okay, you can follow. Is it too fast, too

slow? It's poor English, broken English, it's fine. Is there a book that covers the topics

of the lecture? Hopefully, hopefully, hopefully. No, there isn't. No, I will provide research