A sheaf-theoretic construction of shape space

Presheaves and sheaves, stacks, descent conditions (category-theoretic aspects) (18F20) Derived categories, triangulated categories (18G80) Optimization of shapes other than minimal surfaces (49Q10) Persistent homology and applications, topological data analysis (55N31) Applied homological algebra and category theory in algebraic topology (55U99)

Review text

This article develops a rigorous sheaf-theoretic approach to the mathematical notion of shape space. Classical constructions of shape space, such as Kendall's landmark framework or diffeomorphism-based methods from pattern theory and LDDMM, typically assume topological equivalence of shapes. Such assumptions limit their use in real data settings, especially in biology and anatomy, where shapes are often not homeomorphic. The authors address this limitation by considering constructible subsets of $\mathbb{R}^d$ and assigning to each such subset its persistent homology transform (PHT).\N\NThe main contribution is to show that this assignment defines a homotopy sheaf on the poset category of constructible sets. This allows the PHTs of individual pieces to be glued together to recover the transform of a whole shape. A particularly notable consequence is a nerve lemma for the PHT, which implies that higher-order PHTs of a polyhedral shape can be recovered from zeroth-order PHTs of its covering pieces. Since connected components can be computed efficiently, this yields direct algorithmic benefits.\N\NThe authors also establish stability results for PHTs under controlled homotopies and prove approximation theorems connecting the PHT of a manifold to that of a polyhedral approximation. These findings reinforce the PHT as a robust and computable invariant in shape analysis. The framework requires advanced tools from homological algebra and sheaf theory, and the appendix gives a careful discussion of derived and $\infty$-categorical background, demonstrating the depth of the approach.\N\NThe exposition is mathematically detailed yet well motivated, and it successfully bridges applied topology with modern categorical methods. This contribution will be valuable for researchers in topological data analysis, statistical shape analysis, and related areas of computational geometry.

Cites work

This page was built for publication: A sheaf-theoretic construction of shape space

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6975885)