Subdivision Schemes in Metric Spaces
By: Nira Dyn, Nir Sharon
Potential Business Impact:
Makes computer graphics smoother and more realistic.
We develop a unified framework for nonlinear subdivision schemes on complete metric spaces (CMS). We begin with CMS preliminaries and formalize refinement in CMS, retaining key structural properties, such as locality. We prove a convergence theorem under contractivity and demonstrate its applicability. To address schemes where contractivity is unknown, we introduce two notions of proximity. Our proximity methods relate a nonlinear scheme to another nonlinear scheme with known contractivity, rather than to a linear scheme, as in much of the literature. Specifically, the first type proximity compares the two schemes after a single refinement step and, as in the classical theory, yields convergence from sufficiently dense initial data. The proximity of the second type monitors alignment across all refinement levels and provides strong convergence without density assumptions. We formulate and prove the corresponding theorems, and illustrate them with various examples, such as schemes over metric spaces of compact sets in $\R^n$ and schemes over the Wasserstein space, as well as a geometric Hermite metric space. These results extend subdivision theory beyond Euclidean and manifold-valued data for data in metric spaces.
Similar Papers
Data-intrinsic approximation in metric spaces
Numerical Analysis
Makes computers understand data better with less work.
Discrete Geodesic Calculus in the Space of Sobolev Curves
Numerical Analysis
Helps computers measure how shapes change over time.
Merging Hazy Sets with m-Schemes: A Geometric Approach to Data Visualization
Machine Learning (CS)
Makes complicated data easier to see and understand.