Papers
Topics
Authors
Recent
Search
2000 character limit reached

Ray transform on Sobolev spaces of symmetric tensor fields, I: Higher order Reshetnyak formulas

Published 22 Jun 2021 in math.AP | (2106.11624v1)

Abstract: For an integer $r\ge0$, we prove the $r$th order Reshetnyak formula for the ray transform of rank $m$ symmetric tensor fields on $\mathbb{R}n$. Certain differential operators $A{(m,r,l)}\ (0\le l\le r)$ on the sphere $\mathbb{S}{n-1}$ are main ingredients of the formula. The operators are defined by an algorithm that can be applied for any $r$ although the volume of calculations grows fast with $r$. The algorithm is realized for small values of $r$ and Reshetnyak formulas of orders $0,1,2$ are presented in an explicit form.

Summary

  • The paper establishes explicit higher order Reshetnyak formulas for the ray transform on Sobolev spaces, enabling precise norm equivalence for stability analysis.
  • It introduces novel invariant differential operators on the tangent bundle of the sphere to tackle the challenges of non-ellipticity and complex topology.
  • The work underpins enhanced error estimates and regularization techniques, directly impacting reconstructive algorithms in tensor tomography.

Higher Order Reshetnyak Formulas for the Ray Transform on Sobolev Spaces of Symmetric Tensor Fields

Introduction and Context

This paper establishes a series of higher order Reshetnyak formulas for the ray transform of rank mm symmetric tensor fields on Rn\mathbb{R}^n, extending the foundational analysis for Radon and ray transforms in inverse problems and integral geometry. The Reshetnyak formula provides critical HsH^s-type norm equivalence relations between the initial tensor field and its transform, underlying the stability theory essential for reconstructive algorithms in tensor tomography. Prior to this work, only lower order (specifically, zeroth and first order) Reshetnyak formulas were available for the ray transform, and principally for scalar or low-rank settings; this paper extends those results to arbitrary order and rank, systematically constructing the required machinery of Sobolev spaces on the tangent bundle of the sphere and elucidating the corresponding operator theory.

Sobolev Spaces and Operator Structures

The authors generalize the Sobolev space framework to accommodate the differentiability structure necessary for higher order stability results on TSn−1T S^{n-1}, the tangent bundle of the sphere parameterizing lines in Rn\mathbb{R}^n. The principal technical innovation is the construction of the (generally non-elliptic) operator ΔG\Delta_{\mathcal{G}}, acting on TSn−1T S^{n-1}, analogous to the spherical Laplacian but particularly adapted to the geometry encoded in the ray transform. New Sobolev norms are defined using integer powers of (1+ΔG)(1 + \Delta_{\mathcal{G}}), which provides a route to higher regularity estimates but avoids the analytic obstacles associated with non-ellipticity. Unlike in the hyperplane case (Radon transform), the tangent bundle topology is nontrivial for n>2,4,8n>2,4,8, and so classical spherical harmonic techniques must be replaced with structures attuned to TSn−1T S^{n-1}.

The analysis crucially depends on a hierarchy of invariant algebraic and differential operators: the symmetric multiplication (denoted ii), contraction (jj), inner derivative (dd), and divergence (δ\delta). These act naturally on bundles of symmetric tensors over Sn−1S^{n-1} and admit recursive algebraic composition rules, enabling explicit expression of the operators A(m,r,ℓ)A(m,r,\ell) entering the higher order Reshetnyak formulas as polynomials in these generators.

Main Theoretical Results: Higher Order Reshetnyak Formulas

The central achievement is an explicit formula for the rrth order Reshetnyak identity for the ray transform of rank mm symmetric, solenoidal tensor fields. For each r≥0r\geq 0, the authors construct explicit, self-adjoint, invariant differential operators A(m,r,ℓ)A(m,r,\ell) (with 0≤ℓ≤r0\leq \ell \leq r) acting on the fiber SmT∗Sn−1S^m T^* S^{n-1} such that

∥f∥Ht,sol(r,s)(Rn;SmRn)2=∑ℓ=0r∫0∞∫Sn−1p2t+2ℓ+n−1(1+p2)s−t⟨A(m,r,ℓ)f^(pω),f^(pω)⟩ dω dp\|f\|_{H^{(r,s)}_{t,\text{sol}}(\mathbb{R}^n ; S^m \mathbb{R}^n)}^2 = \sum_{\ell=0}^r \int_{0}^{\infty} \int_{S^{n-1}} p^{2t+2\ell+n-1}(1+p^2)^{s-t} \langle A(m,r,\ell) \hat{f}(p\omega), \hat{f}(p\omega) \rangle \, d\omega \, dp

where the left norm is defined constructively in terms of these operators and the solenoidal (divergence-free) part of ff. The A(m,r,â„“)A(m,r,\ell) are presented algorithmically, their structure governed by intricate recurrence relations (see equations (4.25), (5.18), etc.). For r=0,1,2r=0,1,2, explicit formulae are provided, including concrete coefficient expressions and the appearance of Kronecker contractions and covariant derivatives on the sphere.

The machinery is designed to be invariant under the action of the orthogonal group and isometry operations (translations, rotations), ensuring physical and geometric coherence with the underlying tomography problem. The operators A(m,r,0)A(m,r,0) and A(m,r,r)A(m,r,r) are always positive, directly ensuring norm equivalence and stability.

Technical Implications and Novelty

The explicit construction of higher order norm equivalences for the ray transform on symmetric tensor fields provides, for the first time, an operator-theoretic foundation for quantifying stability of reconstructive procedures at higher regularity levels and for arbitrary tensor ranks. This has direct implications for error analysis and regularization in tensor tomography, including travel time and Doppler tomography.

Notably, while the zeroth order Reshetnyak formula was previously known and isometries for r=0r=0 had been established, the extension to r>0r>0 entails nontrivial computation due to the non-elliptic nature of ΔG\Delta_{\mathcal{G}} and the nontrivial topology of TSn−1T S^{n-1}. The explicit algorithm for constructing the A(m,r,ℓ)A(m,r,\ell) also reveals deep connections to the algebraic combinatorics of tensor contractions and symmetric products.

Further, the treatment naturally incorporates the sharp distinction between tensor field reconstructions: only the solenoidal part (the divergence-free component) of a tensor field is determined by its ray transform, a fundamental point in inverse problems for tensor tomography, reflected in the intrinsic role of the Hodge decomposition.

Future Directions and Open Problems

This work lays the theoretical groundwork for a range of future inquiries:

  • Range characterization for the ray transform of higher rank tensors (announced for a sequel).
  • Extension of higher order Reshetnyak-type formulas to more general manifolds, particularly those with non-flat, non-trivial topology, including magnetic and attenuated settings.
  • Refinement of stability and error propagation estimates for numerical algorithms in tensor tomography, exploiting the quantified higher order norms.
  • Investigation of spectral properties and positivity for the operators A(m,r,â„“)A(m,r,\ell) beyond the explicit cases of small (m,r)(m,r), as conjectures regarding non-negativity and self-adjointness remain partially unresolved for higher ranks and orders.

Additionally, potential exists for applications to microlocal analysis and range completeness studies, given the fully explicit description of the relevant functional analytic framework.

Conclusion

The paper provides a rigorous and comprehensive development of higher order Reshetnyak formulas for the ray transform of symmetric tensor fields, establishing a family of invariant, self-adjoint differential operators governing the Sobolev regularity and norm equivalences central to stability analysis in tensor tomography. The algebraic and analytic techniques developed are expected to serve as foundational tools across integral geometry, inverse problems, and applied tomographic analysis.

Whiteboard

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

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Collections

Sign up for free to add this paper to one or more collections.