Retrieving convex bodies from restricted covariogram functions

Gennadiy Averkov, Gabriele Bianchi

Abstract

The covariogram $g_K(x)$ of a convex body $K \subseteq E^d$ is the function associating each $x \in E^d$ with the volume of the intersection of $K$ with its translate $K + x$. This paper investigates Matheron's conjecture—whether $g_K$ uniquely determines $K$ up to translation and reflection—and specifically explores how much of the covariogram's domain is required for such determination. We demonstrate that for most convex bodies, the covariogram restricted to arbitrarily small subsets of its support, particularly near the boundary or the origin, is sufficient to identify the body's geometric properties.

Introduction

A fundamental problem in convex geometry is determining the extent to which a shape is defined by its self-intersections. Matheron’s conjecture (1986) posits that the covariogram $g_K$ uniquely determines a convex body $K$ up to translation and reflection. While confirmed for polygons and smooth planar bodies, this research shifts the focus toward the "efficiency" of reconstruction: identifying the minimal subsets of the covariogram's support necessary for unique identification. This has significant implications for fields like X-ray crystallography and image analysis, where data acquisition may be restricted to specific regions.

Methodology

The study employs rigorous logical proofs within the framework of convex geometry and Baire category theory. Key methodological components include:

  • Difference Body Analysis: The support of the covariogram is the difference body $DK = K + (-K)$. The geometry of $DK$ is central to understanding how $g_K$ encodes information about $K$.
  • Local Symmetry Arcs: The authors define pairs of boundary arcs that are reflections of each other as "local symmetries." These features act as potential obstacles to unique reconstruction.
  • Baire Category Sense: By proving properties for "most" convex bodies (a residual set), the authors demonstrate that pathological counterexamples are rare and that uniqueness is a generic property.
  • Geometric Tomography: Derivatives of the covariogram are linked to the body's width and projection volumes, facilitating reconstruction from local data.

Key Insights and Geometric Analysis

The research reveals a deep connection between the boundary of the covariogram's support and the original body's structure. For most shapes, the behavior of $g_K$ as the intersection volume approaches zero (near the boundary of $DK$) contains sufficient information for full reconstruction.

Figure 1: (a) A convex body $K$ with local symmetries highlighted; (b) Its difference body $DK$, illustrating the corresponding arcs and midpoints $O$ that encode symmetry information.

A critical insight is that "local symmetries" are the primary bottleneck for uniqueness. If a body lacks these—which is true for the vast majority of convex bodies—reconstruction from restricted data becomes feasible. The paper identifies that for smooth planar bodies, the covariogram restricted to the boundary of its support plus a countable set of midpoints of local symmetry arcs is sufficient for unique determination.

Key Results

  • Genericity in Any Dimension: For most convex bodies in $d \ge 2$, the covariogram restricted to a neighborhood of the boundary of its support uniquely determines the body.
  • Planar Efficiency: In the two-dimensional case, most bodies are also uniquely determined by the covariogram in a neighborhood of the origin (small shifts).
  • Information Duality: The authors prove that information near the origin and information near the boundary are not always redundant, constructing specific examples where one region of data is insufficient while the other succeeds.
  • Extended Uniqueness: The paper extends the class of bodies for which Matheron’s conjecture holds to include all planar convex bodies possessing two nondegenerate boundary arcs that are reflections of each other.

Conclusion and Future Work

This work sharpens our understanding of the covariogram by identifying the minimal data required for geometric reconstruction. While many results are generalized to higher dimensions, the most precise findings regarding local symmetries remain planar. Future research may explore whether knowing the covariogram near the boundary of its support implies a local coincidence of boundaries for smooth bodies in dimensions $d \ge 3$. Notably, the authors have recently extended these findings to confirm Matheron's conjecture for all general planar convex bodies.

Summary updated

Abstract

The covariogram g K ( x ) of a convex body K ⊆ E d is the function which associates to each x ∈ E d the volume of the intersection of K with K + x , where E d denotes the Euclidean d -dimensional space. Matheron (1986) asked whether g K determines K , up to translations and reflections in a point. Positive answers to Matheron's question have been obtained for large classes of planar convex bodies, while for d ≥ 3 there are both positive and negative results. One of the purposes of this paper is to sharpen some of the known results on Matheron's conjecture indicating how much of the covariogram information is needed to get the uniqueness of determination. We indicate some subsets of the support of the covariogram, with arbitrarily small Lebesgue measure, such that the covariogram, restricted to those subsets, identifies certain geometric properties of the body. These results are more precise in the planar case, but some of them, both positive and negative ones, are proved for bodies of any dimension. Moreover some results regard most convex bodies, in the Baire category sense. Another purpose is to extend the class of convex bodies for which Matheron's conjecture is confirmed by including all planar convex bodies possessing two nondegenerate boundary arcs being reflections of each other.
BibTeX
@inproceedings{averkov2007retrieving,
  author = {Gennadiy Averkov and Gabriele Bianchi},
  title = {{Retrieving convex bodies from restricted covariogram functions}},
  booktitle = {Advances},
  year = {2007},
  doi = {10.1239/aap/1189518630},
}
Figure