Vietoris--Rips Shadow for Euclidean Graph Reconstruction
Abstract
The shadow of an abstract simplicial complex $K$ with vertices in $\mathbb R^N$ is a subset of $\mathbb R^N$ defined as the union of the convex hulls of simplices of $K$. The Vietoris--Rips complex of a metric space $(S,d)$ at scale $\beta$ is an abstract simplicial complex whose each $k$-simplex corresponds to $(k+1)$ points of $S$ within diameter $\beta$. In case $S\subset\mathbb R^2$ and $d(a,b)=\|a-b\|$ standard Euclidean, the natural shadow projection of the Vietoris--Rips is already proved to be $1$-connected. We extend the result beyond the standard Euclidean distance on $S\subset\mathbb R^N$ to a family of path-based metrics $d^\varepsilon_{S}$. From the pairwise Euclidean distances of points of $S$, we introduce a family (parametrized by $\varepsilon$) of path-based Vietoris--Rips complexes $R^\varepsilon_\beta(S)$ for a scale $\beta>0$. If $S\subset\mathbb R^2$ is Hausdorff-close to a planar Euclidean graph $G$, we provide quantitative bounds on scales $\beta,\varepsilon$ for the shadow projection map of the Vietoris--Rips of $(S,d^\varepsilon_{S})$ at scale $\beta$ to be $1$-connected. As a novel application, this paper first studies the homotopy-type recovery of $G\subset\mathbb R^N$ using the abstract Vietoris--Rips complex of a Hausdorff-close sample $S$ under the $d^\varepsilon_{S}$ metric. Then, our result on the $1$-connectivity of the shadow projection lends itself to providing also a geometrically close embedding for the reconstruction. Based on the length of the shortest loop and large-scale distortion of the embedding of $G$, we quantify the choice of a suitable sample density $\varepsilon$ and a scale $\beta$ at which the shadow of $R^\varepsilon_\beta(S)$ is homotopy-equivalent and Hausdorff-close to $G$.