Reproduce “SurfGen: Adversarial 3D Shape Synthesis with Explicit Surface Discriminators” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Andrew Luo, Tianqin Li, Wen-Hao Zhang, Tai Sing Lee
SurfGen trains an implicit 3D shape generator by judging extracted surfaces rather than only their underlying representations. It uses differentiable mesh extraction and spherical projections to send adversarial feedback from a spherical CNN discriminator to the generator. On ShapeNet, SurfGen reports the lowest MMD across three tested categories; chair coverage is tied, and car results are less competitive.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.