Dr. Max Fomitchev-Zamilov

And when you approach the handle, the quality of machining gets worse because where that handle meets the vase, you know, there is a ledge and it's not that well polished off. And initially I was just, you know, get rid of that manually. And then I started thinking, well, gee, you know, here's a typical example of a selection effect. You look at a data point, you don't like it, you discard it.

So I spent a lot of time automating this process and I settled on an algorithm that worked like this. I discard every slice of an object. If the root mean square error is outside of the average by the factor of two. And I repeated this process, I think, two or three times.

So I spent a lot of time automating this process and I settled on an algorithm that worked like this. I discard every slice of an object. If the root mean square error is outside of the average by the factor of two. And I repeated this process, I think, two or three times.

So I spent a lot of time automating this process and I settled on an algorithm that worked like this. I discard every slice of an object. If the root mean square error is outside of the average by the factor of two. And I repeated this process, I think, two or three times.

And then that's uniform enough to where it get rid of all of the slices through handles and some of the messed up slices along the top and along the bottom. And without any human intervention. on all the vases. So the algorithm is applying the same set of criteria on, you know, which slices to discard from the analysis without, you know, me having a say in it.

And then that's uniform enough to where it get rid of all of the slices through handles and some of the messed up slices along the top and along the bottom. And without any human intervention. on all the vases. So the algorithm is applying the same set of criteria on, you know, which slices to discard from the analysis without, you know, me having a say in it.

And then that's uniform enough to where it get rid of all of the slices through handles and some of the messed up slices along the top and along the bottom. And without any human intervention. on all the vases. So the algorithm is applying the same set of criteria on, you know, which slices to discard from the analysis without, you know, me having a say in it.

So I can say it's, you know, pretty objective, and somebody can implement a different algorithm, but there is no, like, selection error that arises from human factor. But that's what you need to do. You need to get rid of slices through handles, and some slices along the lip, because as it curves, and some slices at the bottom, you know, because bottom is... on some of these vases is not flat.

So I can say it's, you know, pretty objective, and somebody can implement a different algorithm, but there is no, like, selection error that arises from human factor. But that's what you need to do. You need to get rid of slices through handles, and some slices along the lip, because as it curves, and some slices at the bottom, you know, because bottom is... on some of these vases is not flat.

So I can say it's, you know, pretty objective, and somebody can implement a different algorithm, but there is no, like, selection error that arises from human factor. But that's what you need to do. You need to get rid of slices through handles, and some slices along the lip, because as it curves, and some slices at the bottom, you know, because bottom is... on some of these vases is not flat.

Here's another problem. So it's easy to analyze in a more or less cylindrical objects, but it's not easy to analyze rounded objects with this method. Because as object gets rounded, your slices become oblique. And oblique slices magnify areas disproportionately, kind of like, you know, shadows of a setting sun become elongated and everything becomes like grotesquely enlarged.

Here's another problem. So it's easy to analyze in a more or less cylindrical objects, but it's not easy to analyze rounded objects with this method. Because as object gets rounded, your slices become oblique. And oblique slices magnify areas disproportionately, kind of like, you know, shadows of a setting sun become elongated and everything becomes like grotesquely enlarged.

Here's another problem. So it's easy to analyze in a more or less cylindrical objects, but it's not easy to analyze rounded objects with this method. Because as object gets rounded, your slices become oblique. And oblique slices magnify areas disproportionately, kind of like, you know, shadows of a setting sun become elongated and everything becomes like grotesquely enlarged.

So you get the same thing when you have a rounded bottom vase and you're cutting it obliquely, you just get, you know, noisy, noisy results. Not because, you know, the vase is ill-made. It's just because, you know, the analysis is not suited for oblique cuts. So fortunately, this slice selection algorithm that I discussed in paper, in depth, takes care of that.

So you get the same thing when you have a rounded bottom vase and you're cutting it obliquely, you just get, you know, noisy, noisy results. Not because, you know, the vase is ill-made. It's just because, you know, the analysis is not suited for oblique cuts. So fortunately, this slice selection algorithm that I discussed in paper, in depth, takes care of that.

So you get the same thing when you have a rounded bottom vase and you're cutting it obliquely, you just get, you know, noisy, noisy results. Not because, you know, the vase is ill-made. It's just because, you know, the analysis is not suited for oblique cuts. So fortunately, this slice selection algorithm that I discussed in paper, in depth, takes care of that.

So it discards all slices that are either noisy or going through handles. Yeah, we can go next. Oh, so this is a good scatter. The deviation from the center is the concentricity, right? So all this, I think this is for this vase that I'm holding. So the amazing part here is that the blue dots are the slices for the outer surface and the red dots are the slices for the inner surface.

So it discards all slices that are either noisy or going through handles. Yeah, we can go next. Oh, so this is a good scatter. The deviation from the center is the concentricity, right? So all this, I think this is for this vase that I'm holding. So the amazing part here is that the blue dots are the slices for the outer surface and the red dots are the slices for the inner surface.

So it discards all slices that are either noisy or going through handles. Yeah, we can go next. Oh, so this is a good scatter. The deviation from the center is the concentricity, right? So all this, I think this is for this vase that I'm holding. So the amazing part here is that the blue dots are the slices for the outer surface and the red dots are the slices for the inner surface.

And you see the horizontal axis is in thousands of an inch. And what we see is that the centers of the outer circles are concentric to within two ten thousands of an inch. And outer surface has, sorry, inner surface has slightly larger scatter. And even, you know, when you feel it to the finger, you see that the inner surface is not as well finished.