[= ;307413]
Hey
While we have some work going on in this direction, this is going to be a very large project and probably won’t be available in near future as a lot more research and time is required to get this to level we’d be happy with. Unfortunately, I don’t have a timeframe of when it will be ready to go but once we have something to show we’ll definitely be bringing this up again. There is a Simplygon version of this available already if you have a Simplygon license!
[/]
Thanks for looking into it . I bet it’s going to be pretty large and it’s definitely going to take some time to implement. I’ve written code before to perform subdivisions on a 3d model and I’m curious if this would basically work in reverse, if “de-subdividing” is such a thing, however keeping a focus during process on maintaining prominent vertices. Maybe that would work using some form of pre-calculated edge detection. A following step would be to take that data and preform an averaged process over it to get a pre-map for reconstruction pass that could have an estimated layout for de-subdivided mesh. Maybe next, region out mesh with similar poly density distribution, do a pass to find all primary verts (ie. if cone, tip or maybe a better example, if on a snapping turtles shell, tips of shell), then a pass to reconstruct mesh w/ x% verts relative to density in a given region that was provided from pre-analysis.
I bet rabbit hole could get deep on this one, maybe as well or more so w/ LOD’s but its an interesting topic and one that would save a massive amount of production time and money for lots of developers. Not to mention how quick it would be to go from an art package like 3D into engine… if this problem can be solved, good times ahead 
Thanks again for update and hope all is well!
