Angle‐Based Metric Adaptation for Quadrangulation
Abstract We present an angle‐based metric optimization method for feature‐aware quadrangulation. Unlike existing methods that optimize symmetric positive definite tensors directly, our approach uses triangle angles as the primary optimization variables. This choice has two immediate advantages. First, the constraints for local flatness (a necessary condition for local integrability), sharp features, and corners can be formulated linearly. Second, the positive orientation of the frame (a necessary condition for local injectivity) can be reduced to a simple box constraint on the angles. Moreover, the formulation naturally supports metric customization tasks, including sizing control, conformality, isometric regularization, and singularity specification. The final optimization problem consists of mostly convex terms and can be solved robustly and efficiently. Experiments on CAD models demonstrate improved frame‐field integrability and greater parameterization robustness compared with existing approaches.
Authors
- Hujun Bao (ORCID: https://orcid.org/0000-0002-2662-0334)
- Shuang Wu (ORCID: https://orcid.org/0009-0009-4762-0673)
- Jin Huang (ORCID: https://orcid.org/0000-0002-2549-6810)
- Yijun Xu (ORCID: https://orcid.org/0009-0004-4580-8916)
Institutions
- Zhejiang University (CN)
Publication Details
- Journal
- Computer Graphics Forum
- Published
- 2026-11-13
- DOI
- https://doi.org/10.1111/cgf.70613
- Primary Topic
- Computational Geometry and Mesh Generation
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China