top | item 40572325

(no title)

grondilu | 1 year ago

Please elaborate. Are you thinking of approximating the distance function with an algebraic expression, with algebra itself being the "programming language"?

discuss

order

Philpax|1 year ago

You can represent arbitrary shapes through the composition of SDFs: https://iquilezles.org/articles/distfunctions/

These can be treated as parameterised nodes in a tree, similar to what's happening here. It follows that there may be a possible adaptation of this to SDF composition, such that you can give it a shape and have it produce the SDF nodes + composition required to produce that shape.

Most existing approaches to SDFs with NNs have the NN itself take on the role of the SDF (i.e. given a point, it predicts the distance), so there's a compelling opportunity here to build a system that can produce spatial representations from existing imagery without NN inference at render-time.

I imagine adding the third dimension to the problem makes it much harder, though! I'll have to give the paper a read to determine how coupled to 2D their current approach is.