Bilge Arıkca (Mar 21 2026 at 10:54): Bilge Arıkca (Mar 21 2026 at 10:54):Hii, I'm interested in GSoC project #106 - Automated Surface Skinning. I've been researching some approaches and found algorithms like Voxel-based Shrink-wrapping, Visible Surface Manifold, and Alpha Shapes. I'm not entirely sure which one would be the most suitable for BRL-CAD's existing structure. Does anyone have reccomendations or would it be better if I get in contact with project's potential mentors first?
Sean (Mar 22 2026 at 15:22):Hi @Bilge Arıkca You're already asking in the right place! Welcome. Those are all great and relevant methods with various tradeoffs. Which one to latch onto really will depend on the depth of your familiarity with graphics algorithms in general. Shrink-wrapping in general is a vast topic with extensive academic literature and attempts using different methods.
Personally, I'm keen on trying import into openvdb, exterior-meshing in that space, and then exporting simply because it's very easy to implement -- low risk software engineering. That isn't speaking to runtime or memory implications or how good the end result is, but it's something I've been itching to try for a long time.
There are certainly many advanced alternative methods and it'd be nice if you could research and/or summarize the latest research in proposing a direction. BRL-CAD supports many different structures, but the biggest issue you'll find is adapting an algorithm when the input geometry is hybrid representation, ie.., potentially any format. Not just triangles, not just nurbs, potentially a combination of both plus implicit geometry plus boolean operations and more. What's your experience with representation formats and do you think you can do a quick "recent research" survey of skinning/meshing methods? Siggraph is the golden standard for papers.
Last updated: Mar 31 2026 at 01:22 UTC