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

Okay, so welcome to the Monday lecture. Today we're going to finish the chapter on hand-eye calibration,

which is an important part for interventional procedures.

And the hand-eye calibration basically computes the transformation between the camera and the hand that is carrying the camera,

the two coordinate systems.

And you remember the picture we have seen last time.

We have here the hand coordinate system and the camera coordinate system, which is usually two-dimensional.

And this transformation here is usually fixed, and the whole object here is rotated and translated in space.

So what you should have in mind if we talk about hand-eye calibration is a robot that is carrying an X-ray tube and a detector.

Or another picture we can have in mind is an endoscope, which carries the lens system and the camera,

and the hand of the surgeon who is moving around the endoscope.

Or another example that you can have in mind, and which dealt basically as the motivation for this chapter for us,

is that we are using an ultrasound probe, ultrasound sensor, which generates two-dimensional ultrasound images.

And we have markers attached to the ultrasound sensor, and we want to compute the transformation between the marker coordinate system

and the ultrasound image coordinate system.

And why is it important to have this transform for ultrasound?

Chu. Chu is right? The right pronunciation? Chu? Chu? Chu? No, no, no, no. Once again. Chu? Chu?

So is there a gu at the end? Chu? Okay.

It's important because you want to make a 3D.

3D. You want to know where the image frame is in 3D that you can build up a three-dimensional volume.

So hand-eye calibration is a topic. Then we talked about magnetic navigation as one example.

And in this context, we learned about which type of geometry were the essential matrix,

and the essential matrix includes rotation and translation and correlates the corresponding points into images.

What else did we talk about? 3D ultrasound in general.

And we learned about structure from motion approaches and camera modeling and these things.

What else did we consider in terms of medical interventional applications?

Judith, right? I'm learning. Oh! Judith is remembering this, the half-transform.

Why was this important? In which context did we discuss the half-transform?

You don't remember? What's your name again? Mathias. My brain is sucked out.

Half-transform we used for the shadow detection. You remember to detect lines?

We also extended to detect circles and line artifacts.

What else did we talk about in terms of medical applications, Roman?

Finding homogeneous regions, edge regions using the structure tensor ST.

And then last week we have seen SIFT features, rotation invariant robust features.

We learned about HAWK. What does HAWK stand for?

Mathias, it's dangerous because last year I have observed that my students assumed that I have no clue

what my research assistants are doing while they are giving the lecture here,

but I'm very well informed about the contents, be sure.

And the probability that I will ask a question regarding the lectures that were taught by one of my TAs

is as high as the probability for all the other questions.

You can browse the web and read the blogs, and there is one statement which is really true,

be prepared, he's asking everything. Randomized.

But usually very nice questions. A very nice student shows up, I ask very nice questions.

And the hardness of the questions increases with the type of arrogance you are showing.

If you bring muffins, I mean, will it get worse? For sure, no.

Is there a chance that it will get better? Well, you know, I'm a former Siemens member.

I'm as robust as my algorithms. Bring whatever you can carry.

Just have to know I'm always talking about alcoholic beverages, but I actually do not drink alcoholic beverages,

so bring something, some food or some nice stuff. It's not a disadvantage.

What does Hock stand for, Matthias? Me too. That is not a good excuse. Me too.