Papers
Topics
Authors
Recent
Search
2000 character limit reached

Translation-Invariant Curvature Measures

Updated 27 January 2026
  • Translation Invariant Curvature Measures are finitely additive, translation-invariant functionals that capture localized geometric data on convex bodies.
  • They possess rich algebraic structures as modules over GL(n) and O(n), with key results such as the length ≤ 2 irreducibility conjecture and the Bernig embedding linking to valuations.
  • Applications include Steiner-type volumetric expansions and the classification of angular measures, which bridge convex geometry, integral geometry, and valuation theory.

Translation-invariant curvature measures are finitely additive, translation-invariant, and local-valued functionals on the space of convex bodies in Rn\mathbb{R}^n, encoding localized geometric data such as Federer's curvature measures and derived from the intersection theory of convex geometry, integral geometry, and valuation theory. These measures are equipped with rich algebraic and representation-theoretic structures, particularly as modules over general linear and orthogonal groups, and are foundational for understanding local geometric invariants on convex bodies, polytopes, and submanifolds.

1. Definitions and Structural Properties

A convex body KRnK \subset \mathbb{R}^n is a non-empty compact convex subset. A translation-invariant curvature measure is a map

Φ:K(Rn)M(Rn)\Phi: \mathbb{K}(\mathbb{R}^n) \to \mathcal{M}(\mathbb{R}^n)

assigning to each convex body KK a signed Borel measure Φ(K,)\Phi(K, \cdot), satisfying three essential properties:

  • Locality: Φ(K,β)=Φ(L,β)\Phi(K,\beta) = \Phi(L,\beta) if KU=LUK \cap U = L \cap U for some open UβU \supseteq \beta.
  • Translation-invariance: Φ(K+x,β+x)=Φ(K,β)\Phi(K+x, \beta + x) = \Phi(K,\beta) for all xRnx\in\mathbb{R}^n.
  • Continuity: Weak continuity in the Hausdorff topology: KiKK_i \to K implies Φ(Ki)Φ(K)\Phi(K_i) \Rightarrow \Phi(K).

The collection Curv(Rn)\mathrm{Curv}(\mathbb{R}^n) of all such measures forms a real (or complex) vector space, naturally graded by homogeneity (degree rr) and parity (even/odd under reflection): Φ=Φ0+Φ1++Φn,Curvr=Curvr+Curvr.\Phi = \Phi_0 + \Phi_1 + \cdots + \Phi_n,\qquad \mathrm{Curv}_r = \mathrm{Curv}_r^+ \oplus \mathrm{Curv}_r^-. Prime examples include Federer's curvature measures CrC_r, which are obtained by restricting the rrth support measure Θr\Theta_r to normal cycles, and the Lebesgue measure localization KVol(K)K \mapsto \mathrm{Vol}(K\cap\cdot).

For r=0r=0 and r=nr=n, the structure is especially simple: C0C_0 globalizes to the Euler characteristic; CnC_n is Lebesgue measure. For r=n1r=n-1, curvature measures are parametrized by continuous functions f ⁣:Sn1Rf\colon S^{n-1}\to\mathbb{R} (modulo linear), encoding the classical Minkowski-type functionals (Schuhmacher et al., 20 Jan 2026).

2. Representation Theory and the Length-2\le2 Conjecture

The space Curv(Rn)\mathrm{Curv}(\mathbb{R}^n) admits a natural action by GL(n)\mathrm{GL}(n),

(gΦ)(K,β)=Φ(g1K,g1β),(g\cdot\Phi)(K,\beta) = \Phi(g^{-1}K, g^{-1}\beta),

with the subspace of total-mass-zero measures,

Zr±={ΦCurvr±:Φ(K,Rn)=0  K},Z_r^\pm = \{\Phi \in \mathrm{Curv}_r^\pm : \Phi(K, \mathbb{R}^n) = 0 \; \forall K\},

being GL(n)\mathrm{GL}(n)-invariant.

Conjecture (Length 2\le 2): For each 0rn20\leq r\leq n-2, GL(n)\mathrm{GL}(n) acts irreducibly on Zr±Z_r^\pm. Thus, every Curvr±\mathrm{Curv}_r^\pm has composition series of length at most 2 as GL(n)\mathrm{GL}(n)-module (Schuhmacher et al., 20 Jan 2026).

This conjecture, when coupled with Alesker's irreducibility of translation-invariant valuations, implies a uniform module structure, fundamentally constraining the types of geometric phenomena translation-invariant curvature measures can encode.

3. Proved Cases and Main Technical Tools

For r=0r=0 and r=n2r=n-2, the conjecture is verified:

  • For r=0r=0: The only nontrivial measure is the Euler-kernel, giving a $1$-dimensional irreducible module.
  • For r=n2r=n-2: Injective Bernig embedding B:Zn2Valn1(Rn,Rn)B: Z_{n-2} \to \mathrm{Val}_{n-1}(\mathbb{R}^n, \mathbb{R}^n) relates Zn2Z_{n-2} with vector-valued valuations. Highest-weight theory and precise comparison of K\mathrm{K}-types via (g,K)-module methods establish irreducibility.

The proof methods deploy:

  • Kiderlen–Weil decomposition: Decomposition of measures on polytopes through facewise sums involving cone-valuations;
  • Weil's valuation property: The valuation identity holds for all such measures;
  • Globalization map: glob:Φ[KΦ(K,Rn)]\mathrm{glob}:\Phi \mapsto [K\mapsto\Phi(K,\mathbb{R}^n)] maps curvature measures onto valuations, with kernel ZrZ_r;
  • Bernig embedding: ZrValr+1(Rn,Rn)Z_r\hookrightarrow \mathrm{Val}_{r+1}(\mathbb{R}^n, \mathbb{R}^n), injective and GL(n)\mathrm{GL}(n)-equivariant;
  • (g,K)-module and Casselman–Wallach theory: For precise multiplicity and length bounds.

These approaches reduce the structural problem for curvature measures to explicit and tractable questions in representation theory (Schuhmacher et al., 20 Jan 2026).

4. Classification of Angular Curvature Measures

Angular curvature measures are characterized by the property that their local densities on any polytope are supported on faces and proportional to the corresponding external angles. Any such measure of degree k<n1k < n-1 is determined by an even, 2-homogeneous polynomial on the Plücker image of Grk(Rn)\mathrm{Gr}_k(\mathbb{R}^n), and the correspondence

fΦf,Φf(P,U)=dimF=kf(spanF)γ(F,P)Volk(FU),f \longmapsto \Phi_f, \qquad \Phi_f(P,U) = \sum_{\dim F = k} f(\mathrm{span}F)\gamma(F,P)\mathrm{Vol}_k(F\cap U),

yields a canonical basis. For k=n1k=n-1, angular curvature measures correspond to even smooth functions on Sn1S^{n-1}, with basis described by spherical harmonics (Wannerer, 2018).

Explicitly, for k<n1k<n-1,

dimCurvkang(Rn)=1nk+1(nk)(n+1k+1).\dim\,\mathrm{Curv}_k^{\mathrm{ang}}(\mathbb{R}^n) = \frac{1}{n-k+1}\binom{n}{k}\binom{n+1}{k+1}.

This structure is stable under O(n)\mathrm{O}(n)-action, giving rise to irreducible modules classified by highest weight 2ωnk2\omega_{n-k}.

5. Smooth, Invariant, and Covariant Curvature Measures

The full theory of smooth translation-invariant curvature measures involves integration of specific primitive differential forms over the normal cycle, with the global structure described in three components:

  • Representation Correspondence: Every smooth, translation-invariant, SO(n)\mathrm{SO}(n)-covariant, kk-homogeneous curvature measure with values in an irreducible module Γ[λ]\Gamma_{[\lambda]} is in bijection with a unique primitive form of bidegree (k,n1k)(k, n-1-k) on the sphere bundle.
  • Model and Canonical Forms: Construction of such forms uses model forms in (x,y)(x,y) coordinates, with symmetrization to irreducibles via Young symmetrizers and trace projections. The resulting spaces admit explicit decomposition through the Lefschetz decomposition and symplectic structure on the contact manifold SRnS\mathbb{R}^n.
  • Globalization and Kernel: The globalization map from curvature measures to valuations admits explicit characterization of its kernel in terms of three classes of relations among forms, and its image provides a countable basis of translation-invariant valuations in each degree (Saienko, 2019).

This refined perspective demonstrates that curvature measures are localized versions of valuations, providing more granular control over local geometric content.

6. Characterization of Federer's Curvature Measures and Axiomatic Results

Schuhmacher–Wannerer's structure theorem provides an axiomatic characterization: any SO(n)\mathrm{SO}(n)-invariant translation-invariant curvature measure is necessarily a linear combination of Federer's curvature measures CrC_r. Notably, this result holds without any nonnegativity assumption and follows from Hadwiger's theorem on globalizations combined with the vanishing of SO(n)SO(n)-invariants in the kernel ZZ. This strengthens earlier results by Schneider, which required a positivity assumption (Schuhmacher et al., 20 Jan 2026).

Thus, Federer's measures are uniquely determined among all SO(n)\mathrm{SO}(n)-invariant translation-invariant curvature measures by the basic invariance and locality axioms.

7. Steiner-Type Expansions and Local Formulas

Classical Steiner-type formulas connect curvature measures to volumetric expansions such as

Vol(KtBnβ)=i=0ntniΘi(K,β×Sn1)ωni,\mathrm{Vol}(K \oplus tB^n \cap \beta) = \sum_{i=0}^n t^{n-i} \frac{\Theta_i(K, \beta \times S^{n-1})}{\omega_{n-i}},

with support measures Θr\Theta_r, and yield explicit localizations for intrinsic volumes. The intrinsic volumes ViV_i are recovered as globalizations of the kernels CiC_i in the corresponding component of Curvi\mathrm{Curv}_i. The normal cycle formalism admits explicit invariant differential form representations for curvature measures, unifying geometric, combinatorial, and measure-theoretic viewpoints (Schuhmacher et al., 20 Jan 2026).

Table: Fundamental Structural Aspects

Aspect Description Source (arXiv ID)
Grading Degree (rr) and parity (even/odd under reflection) (Schuhmacher et al., 20 Jan 2026)
GL(n)-action Natural, irreducibility in kernel conjectured (length 2\le 2), verified for r=0r=0, n2n-2 (Schuhmacher et al., 20 Jan 2026)
Axioms Locality, translation-invariance, continuity; valuation property holds (Schuhmacher et al., 20 Jan 2026)
Angularity Classifies via 2-homogeneous polynomials; links to external angles and Plücker coordinates (Wannerer, 2018)
Smooth theory Differential forms on the sphere bundle, Lefschetz decomposition, explicit basis via globalisation (Saienko, 2019)

The study of translation-invariant curvature measures thus bridges geometric measure theory, convex geometry, and modern representation theory, providing a comprehensive local theory underpinning both global valuation invariants and finer angular/geometric analysis.

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Translation Invariant Curvature Measures.