Confirmation of Matheron's conjecture on the covariogram of a planar convex body

Gennadiy Averkov, Gabriele Bianchi

Abstract: The covariogram $g_K$ of a convex body $K$ in $\mathbb{E}^d$ is the function which associates to each $x \in \mathbb{E}^d$ the volume of the intersection of $K$ with $K+x$. In 1986, G. Matheron conjectured that for $d = 2$, the covariogram $g_K$ determines $K$ within the class of all planar convex bodies, up to translations and reflections in a point. This problem is equivalent to challenges in stochastic geometry, probability, and the phase retrieval problem in Fourier analysis. It is also critical for determining the atomic structure of quasicrystals from X-ray diffraction images. This paper confirms Matheron’s conjecture completely.

Introduction

In the field of geometric tomography, a central question is how much information about a shape can be recovered from its various projections or measurements. One such measurement is the covariogram, denoted as $g_K(x)$. For a convex body $K$, the covariogram measures the area (in 2D) or volume (in higher dimensions) of the intersection between $K$ and its translated version $K+x$.

Georges Matheron conjectured in 1986 that this function serves as a unique identifier for any planar convex body, provided we account for trivial symmetries: translation and reflection through a point. While the conjecture was proven for specific cases like polygons or non-smooth bodies over the years, the general case for all planar convex bodies remained one of the most significant open problems in the field until this work by Averkov and Bianchi.

The Geometry of the Covariogram

The difficulty in proving Matheron's conjecture lies in the fact that the covariogram is a non-linear measurement. The problem is deeply linked to the phase retrieval problem in Fourier analysis—reconstructing a signal from the modulus of its Fourier transform—and has direct applications in crystallography, where researchers attempt to reconstruct atomic structures from diffraction patterns.

Methodology: Inscribed Parallelograms and Differential Geometry

The proof provided by Averkov and Bianchi addresses the final and most challenging category of shapes: strictly convex and $C^1$ regular (smooth) bodies. Their approach integrates tools from differential geometry and mathematical analysis.

An inscribed parallelogram $D(x)$ within a convex body $K$, illustrating the geometric relationship between the shift $x$ and the body's boundary.

A key component of the proof involves analyzing the first and second derivatives of the covariogram. The authors demonstrate that the gradient of $g_K(x)$ is intimately related to specific parallelograms inscribed within the body $K$. As shown in the figure above, for a given shift $x$, the geometry of the intersection is defined by the boundary points $p_1, p_2, p_3, p_4$ and their respective normals $u_1, u_2, u_3, u_4$.

Furthermore, the authors utilize the Monge–Ampère equation, a non-linear partial differential equation. They show that the determinant of the Hessian matrix of the covariogram acts as a "symmetry detector." Specifically, if the determinant is constant (specifically -1) for all $x$, the body must be centrally symmetric. This allows the researchers to distinguish between symmetric and non-symmetric bodies based solely on the covariogram.

Key Results

The primary contribution of this research is the formal proof of Theorem 1.1: Every planar convex body is uniquely determined within the class of all planar convex bodies by its covariogram, up to translations and point reflections.

Beyond the main theorem, the paper resolves three related problems (P1-P3) in stochastic geometry and Fourier analysis. It confirms that a 2D body can be uniquely determined from:

  1. Its chord-length distribution.
  2. The distribution of the difference of random variables uniformly distributed in $K$.
  3. The modulus of its Fourier transform.

Implications and Future Work

While this work provides a definitive answer for two-dimensional space, the problem remains complex in higher dimensions.

  • The 3D Case: Matheron’s conjecture is still unproven for general convex bodies in three dimensions, although it has been verified for 3D polytopes.
  • Higher Dimensions: The conjecture is known to be false for $d \ge 4$, where counterexamples exist—distinct, non-equivalent bodies that yield identical covariograms.
  • Non-Convexity: The uniqueness property fails if the convexity requirement is removed, as there are many examples of non-convex sets with identical covariograms.

This paper effectively "closes the book" on the planar case, providing a rigorous foundation for future explorations into the 3D mystery and the specific conditions under which higher-dimensional bodies might still be uniquely determined.

Summary updated

Abstract

The covariogram g_K of a convex body K in \mathbb E^d is the function which associates to each x\in \mathbb E^d the volume of the intersection of K with K+x . In 1986 G. Matheron conjectured that for d = 2 the covariogram g_K determines K within the class of all planar convex bodies, up to translations and reflections in a point. This problem is equivalent to some problems in stochastic geometry and probability as well as to a particular case of the phase retrieval problem in Fourier analysis. It is also relevant for the inverse problem of determining the atomic structure of a quasicrystal from its X-ray diffraction image. In this paper we confirm Matheron’s conjecture completely. This problem is equivalent to some problems in stochastic geometry and probability as well as to a particular case of the phase retrieval problem in Fourier analysis. It is also relevant for the inverse problem of determining the atomic structure of a quasicrystal from its X-ray diffraction image. In this paper we confirm Matheron's conjecture completely.
BibTeX
@article{averkov2009confirmation,
  author = {Gennadiy Averkov and Gabriele Bianchi},
  title = {{Confirmation of Matheron's conjecture on the covariogram of a planar convex body}},
  journal = {Journal of the European Mathematical Society},
  year = {2009},
  doi = {10.4171/jems/179},
}
Figure