I hope you had a nice week with Matthias last week.

He will also replace me next week.

Yeah, where it's mainly continuation of this week where we talk about normalization and

feature extraction.

Okay, other questions to the last week?

I make a little, little repetition in form of the quiz again.

And let's see, I hope everyone has that already.

Okay, let's close that.

And how can we make it bigger here?

Okay, so the first question basically is you have now a signal and if it's somehow with

a nonlinear noise, could be a convolutive noise or just a multiplicative noise.

And a nice theory is homomorphic filtering to deal with such a problem.

And homomorphic filtering basically translates the problem to a linear problem.

And here you just shall order now the three steps here in the correct order.

Should be quite straightforward, I hope.

Everyone needs the QR code again.

Yeah?

Okay.

Good.

Okay, 15, not so many people here today.

So I guess we can already see what you got.

Okay, so the most people have that, right?

So of course you have to first transform the signal into a linear form.

So in case of multiplicative, you take the logarithm and then you can apply a linear

filter on top of it and afterwards you have to revert of course the system again and take

the inverse system.

So the most of you got that right.

That's nice.

Let's quickly go to the next one.

I think some of you will need again the code.

But here, so two of the answers here are correct.

So you have to cross which is which.

And I know these are only names for, but these are specific definitions now.

Erosion, deletion, and opening and closing.

And I mean this is something you have to just to learn.

The name is what is behind that and these are mathematical, morphological operations

and you can do quite some nice things with it.

And here I just want that you know these definitions.

For those who came now a little bit late, maybe we can, do you have to?

So okay, maybe again for those people arriving, just arrived.

Here's the QR code again.

Okay, otherwise it's 7DS2.

Okay, 60 people, that seems to be fine.

Okay, yeah, that is quite nice.

So dilation and erosion, the most of you had correct.

So dilation fills small holes and also gets, the object gets a little bit larger.

And in erosion, you remove small structures and the object also gets a little bit smaller.

And because you have these, yeah, that the objects get smaller, you maybe also want to

revert this process.

And this can be done if you combine now erosion and dilation, for example, in opening or in