Previous | Next --- Slide 32 of 70
Back to Lecture Thumbnails
ceviri

Is it not possible to do some sort of multivariate taylory thing to get a polynomial representation?

Is it even computationally useful to do something like that?

Shell

Could you do something where you split up the more complex shapes into smaller surfaces and then somehow combine the equations together? Kind of like functions, where if you apply g(x) to f(x) then you can get a graph that looks like a combination of the two graphs?

L100magikarp

@Shell I guess combining smaller, locally-polynomial surfaces into more complex ones is the intuition of the B-splines except that you combine them piecewise rather than as compositions of functions (f(g(x)).

kallico

Assume a spherical cow in a vacuum. QED

keenan

@ceviri You can fit polynomials to surfaces, just takes some serious optimization (and often isn't guaranteed to work). More generally there are lots of techniques for constructing implicit surfaces from polygon soup, like this paper: https://people.engr.tamu.edu/schaefer/teaching/689_Fall2006/Shen-2004-IAI.pdf

implicit surface generated from polygon soup