Reproduce “QEMesh: Employing A Quadric Error Metrics-Based Representation for Mesh Generation” 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
Jiaqi Li, Ruowei Wang, Yu Liu, Qijun Zhao
QEMesh generates 3D meshes from local surface descriptions—points, normals, and Quadric Error Metric matrices—using latent diffusion and an occupancy decoder to select voxel cells. It adapts PoNQ rather than introducing it. On 64³ ShapeNet airplanes, it reports MMD-CD 3.751 versus GenUDC 3.960, alongside 100% watertightness and 0% self-intersections; the evaluation covers five categories without establishing broader validity.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.