Simplifying Triangle Meshes in the Wild

Autor: Liu, Hsueh-Ti Derek, Zhang, Xiaoting, Yuksel, Cem
Rok vydání: 2024
Předmět:
Druh dokumentu: Working Paper
Popis: This paper introduces a fast and robust method for simplifying surface triangle meshes in the wild while maintaining high visual quality. While previous methods achieve excellent results on manifold meshes by using the quadric error metric, they suffer from producing high-quality outputs for user-created meshes, which often contain non-manifold elements and multiple connected components. In this work, we begin by outlining the pitfalls of existing mesh simplification techniques and highlighting the discrepancy in their formulations with existing mesh data. We then propose a method for simplifying these (non-manifold) triangle meshes, while maintaining quality comparable to the existing methods for manifold inputs. Our key idea is to reformulate mesh simplification as a problem of decimating simplicial 2-complexes. This involves a novel construction to turn a triangle soup into a simplicial 2-complex, followed by iteratively collapsing 1-simplices (vertex pairs) with our modified quadric error metric tailored for topology changes. Besides, we also tackle textured mesh simplification. Instead of following existing strategies to preserve mesh UVs, we propose a novel perspective that only focuses on preserving texture colors defined on the surface, regardless of the layout in the texture UV space. This leads to a more robust method for textured mesh simplification that is free from the texture bleeding artifact. Our mesh simplification enables level-of-detail algorithms to operate on arbitrary triangle meshes in the wild. We demonstrate improvements over prior techniques through extensive qualitative and quantitative evaluations, along with user studies.
Databáze: arXiv