Pith. sign in

REVIEW 4 major objections 4 minor 36 references

Numerical reconstruction of radiative sources in an absorbing and non-diffusing scattering medium in two dimensions

T0 review · 4 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A Fourier-mode splitting reduces source reconstruction in an absorbing, anisotropic scattering medium to a non-scattering inversion plus Poisson solves, and is exact when the scattering kernel has finite angular Fourier content.

desk verdict A concrete numerical realization of the Fourier-mode inverse transport method, worth refereeing despite missing noise analysis and a load-bearing citation to an unpublished companion paper. read the letter →

arxiv 1908.09133 v1 pith:NBLS6QT2 submitted 2019-08-24 math.NA cs.NA

Hiroshi Fujiwara , Kamran Sadiq , Alexandru Tamasan This is my paper · ORCID
classification math.NAcs.NA MSC 65N2130E20
keywords transportequationinverseproblemsnumericalsourcereconstructionattenuatedX-raytransformRadonA-analyticmapsHilbertopticalmolecularimaging
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper presents a numerical algorithm that recovers the internal radiative source $q$ from outgoing boundary measurements in a two-dimensional absorbing and anisotropic scattering medium, in the regime where scattering is neither negligible nor strong enough for the diffusion approximation. The central idea is to split the scattering kernel at a finite angular Fourier order $M$, which separates the problem into a non-scattering tail inversion and a finite chain of elliptic boundary value problems. For scattering kernels with genuinely finite Fourier content, the paper argues the reconstruction is exact and stable. For general kernels such as the Henyey-Greenstein kernel, the truncation error in the boundary data can be made smaller than the noise level, and the numerical experiments reconstruct piecewise constant sources quantitatively. This matters for optical molecular imaging, where realistic scattering is significant but not diffusive.

What carries the argument

The machinery is the Fourier decomposition of the transport solution and scattering kernel in the angular variable, which converts the transport equation into an infinite elliptic system. When the scattering kernel has only $M$ nonzero angular Fourier modes, modes above $M$ satisfy the scattering-free system, which is solved by the Cauchy-type integral formula that recovers interior values from boundary values after conjugation by the integrating factor $e^{-h[\mu_t]}$, where $h[\mu_t]$ is the attenuated X-ray phase defined in the paper and chosen so that all negative Fourier modes vanish. The lower modes are then recovered recursively by solving Dirichlet problems for the Poisson equation, and the source $q$ is read off from the zero-mode equation. The algorithm also includes a locally optimal truncation criterion: choose $M$ to minimize the imaginary part of the reconstructed $I_0$, an indicator computable without knowing $q$.

What would settle it

Run the algorithm on boundary data generated from a known source with controlled Gaussian noise added at levels $\delta = 10^{-2}, 10^{-3}, 10^{-4}$; if the reconstruction error does not stay bounded as the mesh is refined for a fixed $\delta$, or if no truncation order $M$ yields an acceptable source, the exact-data premise is falsified. A simpler check is to compare the predicted kernel-truncation data error $g^{2M+2}/(1-g^2)$ with the actual measured boundary error for a range of $g$ and $M$ in the paper's settings.

Watch

Extended reading notes

Core claim

The central claim is that any finite Fourier content in the angular variable splits the inverse transport problem into a non-scattering part and a boundary value problem for a finite elliptic system, and that this split can be implemented numerically. For scattering kernels of polynomial type in the angular variable, the algorithm is claimed to recover the source exactly and stably, with no smallness assumption on the scattering modes. For general kernels, the paper gives an error estimate showing that the mismatch between the measured outflow and the outflow produced by the $M$-truncated kernel is bounded by a constant times the $L^2$ difference of the kernels; choosing $M$ so that this falls below the noise level yields a minimum-residual reconstruction. The numerical experiments with strongly scattering parameters (mean free path $1/5$, anisotropy parameter $g=1/2$) reconstruct piecewise constant sources with quantitatively correct values and support, using a truncation order selected by minimizing the imaginary part of the zero Fourier mode, a criterion that does not need the unknown source.

Load-bearing premise

The reconstruction is demonstrated only on noiseless, numerically generated boundary data, and the paper provides no regularization or analysis of how the discretized Cauchy integral and recursive Poisson solves behave under measurement noise, even though the inverse problem is ill-posed.

Editorial extensions

If this is right

  • For any scattering kernel with finitely many nonzero angular Fourier modes, the algorithm recovers the source exactly from noiseless boundary data, without any smallness assumption on the scattering strength.
  • For the two-dimensional Henyey-Greenstein kernel, the error estimate yields explicit decay $g^{2M+2}/(1-g^2)$; selecting $M$ so this is below the noise level makes the reconstructed source's boundary data indistinguishable from the exact data within that noise.
  • The truncation criterion based on the imaginary part of $I_0$ selects a reasonable $M$ in experiments and tracks the actual reconstruction error, and it requires no knowledge of the unknown source.
  • The method operates in a strongly scattering regime (average of ten scatterings across the unit disc) where the diffusion approximation fails, and still produces quantitative reconstructions of piecewise constant sources.
  • Reconstructions in the experiments run in minutes on a multicore workstation, indicating practical feasibility for optical molecular imaging parameters.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the same angular-Fourier splitting should extend to three-dimensional transport with axial symmetry, replacing the unit circle by the sphere and the Cauchy formula by a spherical harmonics analogue; this is a testable extension the paper does not address.
  • Editorial inference: the minimum-imaginary-part criterion is a general model-selection heuristic: whenever the true quantity is real, a reconstruction parameter can be tuned by minimizing the spurious imaginary component, independent of ground truth.
  • Editorial inference: because the experiments use noiseless simulated data and the error estimate covers only kernel truncation, a natural next test is to add measurement noise and see how the optimal $M$ and the reconstruction error degrade; the paper's method as stated does not include regularization for that case.
  • Editorial inference: the recursive Poisson solves propagate information from high to low Fourier modes, so errors in the tail inversion should appear most strongly in the finer spatial features of $q$; comparing reconstructions at different $S$ and mesh resolutions would reveal this error propagation.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The manuscript presents a numerical algorithm for reconstructing a compactly supported source q in the stationary linear transport equation in a bounded convex two-dimensional domain, from knowledge of the angularly resolved outgoing boundary radiation. The medium is allowed to have anisotropic scattering that is neither negligible nor strong enough for the diffusion approximation. The method expands the intensity and the scattering kernel in angular Fourier modes; for kernels with finitely many angular modes it reduces the problem to a non-scattering Cauchy-type problem for the high modes plus recursive Poisson solves for the low modes, after which q is recovered from the zeroth and first Fourier modes. For the Henyey-Greenstein kernel, an M-th order truncation of the kernel is used, and a heuristic criterion based on the imaginary part of the reconstructed I0 is proposed to choose M. Two numerical experiments, using boundary data generated by an independent high-resolution forward solver and a coarser reconstruction mesh, produce quantitatively accurate reconstructions of piecewise constant sources. The exactness and stability for polynomial-type scattering kernels are stated as consequences of the authors' companion manuscript [14], which is listed as under review.

Significance. If the claims hold, this is a useful first numerical realization of the Bukhgeim-type Fourier reconstruction method for strongly anisotropic scattering, and the experiments are well designed: they avoid inverse crime by using a piecewise-constant upwind forward solver on a very fine mesh and a separate, coarser reconstruction mesh, and they report quantitative errors as well as computational cost. The paper also contains a clean treatment of the Hilbert transform in Lemma 1 and a detailed step-by-step algorithm that should be reproducible. However, the paper's central exactness/stability claim is not self-contained, and the absence of any noise-contaminated experiment or regularization analysis is a substantial gap for an inverse problem that the paper itself identifies as ill-posed. The reconstruction's behavior under measurement noise is the weakest part of an otherwise competent numerical study.

major comments (4)
  1. [Abstract; Section 3; [14, Corollary 6.1]] The exactness and stability assertion for scattering kernels of finite Fourier content is the load-bearing theoretical foundation of the paper, but it is not proved here: the manuscript cites only [14, Corollary 6.1], a companion paper listed as under review. The present paper cannot be fully evaluated without knowing the regularity conditions, the precise sense of convergence, and the stability estimate behind that corollary. Please state the relevant theorem with its hypotheses, include a proof or a detailed proof sketch, and explain how the discretized Steps 5-10 are intended to converge to that infinite-dimensional result.
  2. [Section 3; Section 5; Steps 5, 9, 10] The inverse source problem is explicitly described as ill-posed in the Introduction, yet the algorithm contains no regularization step and all experiments in Section 5 use noiseless, numerically generated data. Proposition 1 bounds the data error caused by kernel truncation, not the effect of measurement noise, and the 'minimum residual' interpretation in Section 3 is not realized as an actual minimization. Steps 5, 9, and 10 involve a Cauchy-type boundary integral, recursive Poisson solves, and differentiation of a finite-element function to recover q, all of which can amplify high-frequency noise. Please add a stability estimate with respect to data noise, or at minimum numerical experiments with realistic noise levels, and state clearly whether and how the choices of M, S, and the mesh act as regularization.
  3. [Section 4.4; Experiment 1; Figure 6] The proposed optimality criterion (18) is not validated by the paper's own experiment. In Experiment 1, Figure 6 shows that the pseudo-error (20) is minimized at M = 3 while the imaginary-part criterion (18) is minimized at M = 6; the text even says 'Both errors take minimum around M = 5' before giving those conflicting values. Since the criterion is presented as a practical, source-independent substitute for the true error, this mismatch needs explanation and further validation, including under noise, before the claim that the algorithm includes a feasible optimality criterion is supported.
  4. [Section 3; Eq. (13); Step 9] The continuous derivation argues by induction that I_m^{(M)} in H^1(D) because the right-hand side of (13a) lies in H^{-1}(D). In the discrete realization, however, each I_{m+1} and I_{m+2} is a P1 finite-element function whose derivative is piecewise constant and discontinuous across triangle edges, and the variational form in Step 9 evaluates the right-hand side cellwise. No convergence or stability analysis is given for this recursive finite-element scheme as m decreases, nor for the boundary interpolation in Step 8. This is directly relevant to whether the exactness claim for polynomial-type kernels applies to the algorithm as implemented; please add an error analysis or at least a systematic mesh-refinement study.
minor comments (4)
  1. [Section 4.1] There is a typo in Lemma 1: 'defied' should be 'defined'.
  2. [Section 5, Experiment 1] The sentence 'Both errors take minimum around M = 5' contradicts the following sentence and Figure 6, which state minima at M = 6 for the imaginary part and M = 3 for the pseudo-error; please correct the text.
  3. [Section 5, Experiment 1] Please state explicitly the number K of boundary measurement points in Experiment 1 (Experiment 2 gives K = 3000 and N = 360, but Experiment 1 only mentions 360 velocity intervals and 100 sampling points for the integral transforms).
  4. [References [13], [14]] Reference [13] is described only as 'accepted' with no year or venue, and [14] is under review; please update these entries or provide fuller bibliographic information so that the numerical method and the theoretical foundation can be located.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the algorithm is derived from an explicit elliptic-system reduction and validated on independently generated forward data; self-citations are supporting lemmas, not the quantity being predicted.

full rationale

The reconstruction chain is self-contained at the level of the derivation: Section 3 reduces the transport equation (1) with a finite-Fourier kernel to the elliptic system (12), transforms the tail modes via the explicit integrating-factor relations (10)-(11), applies the Cauchy-type formula (6), and then recovers I_M, I_{M-1}, ..., I_0 by solving the Poisson problems (13); the source is finally read off from (12a). None of these steps uses q as an input: q appears only in the final algebraic relation, while all boundary and interior quantities are computed from I_measure and the known coefficients. The numerical experiments are also not circular: the boundary data are produced by a different forward solver ([13], with a much finer mesh and 360 velocity directions), the reconstruction mesh is generated without information about the support of q, and the algorithm never invokes the true source. Self-citations to [14], [32], and [33] are to the authors' prior mathematical results (integrating factor construction, range theorems, stability); these are used as lemmas with their own assumptions, and the finite-Fourier splitting is independently credited to [25]. Proposition 1 is a forward-model error estimate for kernel truncation; it does not hide the measured data inside the model. The truncation criterion (18) depends only on the imaginary part of the recovered I0 and is explicitly independent of the unknown source; even if it is suboptimal in Experiment 1, that is a performance issue, not a circular reduction. The lack of a noise analysis is a substantive limitation but not a circularity.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The algorithm rests on standard transport well-posedness, Bukhgeim's A-analytic theory as developed in the authors' prior papers, and on the yet-unpublished stability result [14]. No new physical entities are postulated. The main free parameters are numerical truncation orders, not physical constants.

free parameters (2)
  • truncation order M = 6 (Exp 1), 8 (Exp 2), 40 (Exp 2 alternative)
    Chosen by minimizing the imaginary part of the reconstructed zeroth Fourier mode I0 (Eq. 18), a data-dependent heuristic. In Experiment 1 the true pseudo-error is minimized at M=3 while the criterion picks M=6, so the choice is not tied to the actual reconstruction error.
  • angular mode truncation S = 128
    Chosen as 'sufficiently large' (S >= M+3) without a convergence study; all infinite sums for J_m and I_m are truncated at S.
assumptions (5)
  • standard math The forward transport problem (1) has a unique solution in the stated L2-based space for bounded nonnegative coefficients.
    Invoked in Section 1 and used to justify the Fourier series representation (3); well-posedness cited from [10] and [35].
  • standard math The transformation (10)-(11) with integrating factor h (8) maps solutions of the attenuated system (7) bijectively to the Bukhgeim system (5).
    Stated in Section 2 from [32, Lemma 4.1] and used in Section 3 to reduce the tail system (12c) to (5).
  • standard math The Cauchy-like integral formula (6) recovers interior values of solutions of (5) from boundary data.
    Used in Step 5; originates in Bukhgeim [7].
  • domain assumption For polynomial-type scattering kernels, the recursive Poisson solves (13) produce H1 solutions with stable propagation of regularity from M down to 0.
    Regularity assertions in Section 3 are stated without proof, and exactness/stability is cited to [14, Corollary 6.1], an unreviewed manuscript by the same authors.
  • domain assumption Truncating the Henyey-Greenstein kernel at order M perturbs the boundary data by at most C * g^(M+1)/sqrt(1-g^2), which is below the actual noise level for the chosen M.
    The estimate (14) and the Parseval calculation in Section 3 justify the truncation, but the actual noise level in the experiments is not specified, since the data are noiseless.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Numerical reconstruction of radiative sources in an absorbing and non-diffusing scattering medium in two dimensions." pith.science (2026). https://pith.science/paper/NBLS6QT2

@misc{pith2026190809133,
  author       = {Pith},
  title        = {Pith review of: Numerical reconstruction of radiative sources in an absorbing and non-diffusing scattering medium in two dimensions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NBLS6QT2}},
  note         = {Machine review of arXiv:1908.09133}
}
read the original abstract

We consider the two dimensional quantitative imaging problem of recovering a radiative source inside an absorbing and scattering medium from knowledge of the outgoing radiation measured at the boundary. The medium has an anisotropic scattering property that is neither negligible nor large enough for the diffusion approximation to hold. We present the numerical realization of the authors' recently proposed reconstruction method. For scattering kernels of finite Fourier content in the angular variable, the solution is exact. The feasibility of the proposed algorithms is demonstrated in several numerical experiments, including simulated scenarios for parameters meaningful in optical molecular imaging.

Figures

Figures reproduced from arXiv: 1908.09133 by the authors.

Figure 1
Figure 1. Locations of inclusions in numerical examples. The medium is relatively strongly absorbing inside the dotted balls, while the source q(x) is located in the gray regions. The source, to be reconstructed, is (19) q(x) =    2, in R; 1, in B2; 0, otherwise. The absorption coefficient µa(x) is given by µa(x) =    2, in B1; 1, in B2; 0.1, otherwise. To generate the boundary data, we solve the forward problem usi… view at source ↗
Figure 2
Figure 2. Boundary measurement I(ζ, ξ)|∂D×S1 obtained by the numerical computation of the forward problem in the unit disc (in grey). For ζ ∈ ∂D (indicated by ×), the red curve is { ζ + 2I(ζ, ξ)ξ ; ξ ∈ S 1 } (on the left). The right figure is a magni￾fication of the curve at ζ = (1, 0). ξ ⊥ O ζ ζ · ξ ⊥ 0 π/ π 3π/ 2π -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 arg ξ ζ⋅ξ 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 [PITH_FULL_IMAGE:figu… view at source ↗
Figure 3
Figure 3. Left : Projection of the measurement data I(ζ, ξ)|Γ+ . The red arrows, I(ζ, ξ) with ξ = (1, 0), are projected to the plane with Arg ξ ⊥ = π/2. Right : Projection of I|Γ+ corresponding to [PITH_FULL_IMAGE:figures/full_fig_p014_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Reconstructed source with M = 6 [PITH_FULL_IMAGE:figures/full_fig_p015_4.png]
Figure 5
Figure 5. Figure 5: Section of the reconstructed source along the dotted lines in [PITH_FULL_IMAGE:figures/full_fig_p016_5.png]
Figure 6
Figure 6. Figure 6: The truncation parameter M and pseudo-errors in re￾constructed source (+, left axis) and corresponding imaginary part (×, right axis) in Experiment 1. Experiment 2. Let us consider the situation that µa is given by the modified Shepp-Logan phantom [36], which occupies …
Figure 7
Figure 7. Figure 7: Discontinuous interface of µa (solid curves) in the Shepp-Logan phantom, and the locations of support of the source q (gray) 0 π/ π 3π/ 2π -1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1 arg ξ ζ⋅ξ 0 0.05 0.1 0.15 0.2 0.25 0.3 [PITH_FULL_IMAGE:figures/full_fig_p017_7.png]
Figure 8
Figure 8. Figure 8: Measurement data on ∂D is shown by red curves, and the ellipse domain D is filled by gray (on the left), projection of measurement data (on the right) The boundary measurement is generated by solving the forward problem with 1, 554, 282 triangles and 360 velocity direc…
Figure 9
Figure 9. Figure 9: Numerical reconstruction of the source q(x) (real part) on D with M = 8 in Experiment 2. -0.5 0 0.5 1 1.5 2 2.5 3 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 q(x) x1 along x2=0.000 exact -0.5 0 0.5 1 1.5 2 2.5 3 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 q(x) x1 along x2=-0.605 exact [PITH_FULL_I…
Figure 10
Figure 10. Figure 10: Cross sections of the reconstructed source q(x) with M = 8 on the dotted lines in [PITH_FULL_IMAGE:figures/full_fig_p018_10.png]
Figure 11
Figure 11. Figure 11: The truncation parameter M and pseudo-errors in reconstructed source (+, left axis) and corresponding imaginary part (×, right axis). -0.5 0 0.5 1 1.5 2 2.5 3 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 q(x) x1 along x2=0.000 exact -0.5 0 0.5 1 1.5 2 2.5 3 -0.6 -0.4 -0.2 0 0.2 0.4 0…
Figure 12
Figure 12. Figure 12: Cross sections of the reconstructed source q(x) with M = 40 on the dotted lines in [PITH_FULL_IMAGE:figures/full_fig_p019_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 35 canonical work pages

  1. [14]

    Fujiwara, K

    H. Fujiwara, K. Sadiq, and A. Tamasan, A Fourier approach to the inverse source problem in an absorbing and non-weakly scattering medium . under review

    work page

  2. [1]

    D. S. Anikonov, A. E. Kovtanyuk, and I. V. Prokhorov, Transport equation and tomog- raphy, vol. 30 of Inverse and Ill-posed Problems Series, VSP, Utrecht, 2002

    work page 2002

  3. [2]

    E. V. Arbuzov, A. L. Bukhge ˘ım, and S. G. Kazantsev , Two-dimensional tomography problems and the theory of A-analytic functions [translation of algebra, geometry, analysis and mathematical physics (russian) (novosibirsk, 1996), 6–20, 189, Izdat. Ross. Akad. Nauk Sibirsk. Otdel. Inst. Mat., Novosibirsk, 1997] , Siberian Adv. Math., 8 (1998), pp. 1–20

    work page 1998

  4. [3]

    S. R. Arridge , Optical tomography in medical imaging , Inverse Problems, 15 (1999), pp. R41–R93

    1999

  5. [4]

    Bal , On the attenuated Radon transform with full and partial measurements , Inverse Problems, 20 (2004), pp

    G. Bal , On the attenuated Radon transform with full and partial measurements , Inverse Problems, 20 (2004), pp. 399–418. 20 H. FUJIWARA, K. SADIQ, AND A. TAMASAN

    work page 2004

  6. [5]

    Bal and A

    G. Bal and A. Tamasan , Inverse source problems in transport equations , SIAM J. Math. Anal., 39 (2007), pp. 57–76

    work page 2007

  7. [6]

    Boman and J.-O

    J. Boman and J.-O. Str ¨omberg, Novikov’s inversion formula for the attenuated Radon transform—a new approach, J. Geom. Anal., 14 (2004), pp. 185–198

    work page 2004

  8. [7]

    A. L. Bukhgeim, Inversion formulas in inverse problems , in Linear Operators and Ill-Posed Problems by M. M. Lavrentiev and L. Ya. Savalev, Plenum, New York, (1995), pp. 323–378

    work page 1995

Show all 36 references
  1. [8]

    Choulli and P

    M. Choulli and P. Stefanov , Inverse scattering and inverse boundary value problems for the linear Boltzmann equation, Comm. Partial Differential Equations, 21 (1996), pp. 763–785

    1996

  2. [9]

    Choulli and P

    M. Choulli and P. Stefanov, An inverse boundary value problem for the stationary trans- port equation, Osaka J. Math., 36 (1999), pp. 87–104

    1999

  3. [10]

    Dautray and J.-L

    R. Dautray and J.-L. Lions, Mathematical analysis and numerical methods for science and technology. Vol. 4, Springer-Verlag, Berlin, 1990. Integral equations and numerical methods, With the collaboration of Michel Artola, Philippe B´ enilan, Michel Bernadou, Michel Cessenat, J...

    1990

  4. [11]

    Ern and J.-L

    A. Ern and J.-L. Guermond , Theory and practice of finite elements , vol. 159 of Applied Mathematical Sciences, Springer-Verlag, New York, 2004

    2004

  5. [12]

    D. V. Finch , The attenuated x-ray transform: recent developments , in Inside out: inverse problems and applications, vol. 47 of Math. Sci. Res. Inst. Publ., Cambridge Univ. Press, Cambridge, 2003, pp. 47–66

    2003

  6. [13]

    Fujiwara, Piecewise constant upwind approximations to the stationary radiative transport equation

    H. Fujiwara, Piecewise constant upwind approximations to the stationary radiative transport equation. accepted in Proceedings of International Conference Continuum Mechanics Focusing on Singularities 2018

    2018

  7. [15]

    Hecht, New development in freefem++ , J

    F. Hecht, New development in freefem++ , J. Numer. Math., 20 (2012), pp. 251–265

    2012

  8. [16]

    Helgason, The Radon transform, vol

    S. Helgason, The Radon transform, vol. 5 of Progress in Mathematics, Birkh¨ auser, Boston, Mass., 1980

    1980

  9. [17]

    Hubenthal, An inverse source problem in radiative transfer with partial data , Inverse Problems, 27 (2011), pp

    M. Hubenthal, An inverse source problem in radiative transfer with partial data , Inverse Problems, 27 (2011), pp. 125009, 22

    2011

  10. [18]

    Kuchment, The Radon transform and medical imaging , vol

    P. Kuchment, The Radon transform and medical imaging , vol. 85 of CBMS-NSF Regional Conference Series in Applied Mathematics, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 2014

    2014

  11. [19]

    L. A. Kunyansky , A new SPECT reconstruction algorithm based on the Novikov explicit inversion formula, Inverse Problems, 17 (2001), pp. 293–306

    2001

  12. [20]

    E. W. Larsen, The inverse source problem in radiative transfer, J. Quant. Spectrosc. Radiat. Transfer, 15 (1975), pp. 1–5

    1975

  13. [21]

    N. J. McCormick and R. Sanchez , Solutions to an inverse problem in radiative transfer with polarization–II, J. Quant. Spectrosc. Radiat. Transfer, 30 (1983), pp. 527 – 535

    1983

  14. [22]

    Mokhtar-Kharroubi, Mathematical topics in neutron transport theory , Series on Ad- vances in Mathematics for Applied Sciences, World Scientific, Singapore, 1997

    M. Mokhtar-Kharroubi, Mathematical topics in neutron transport theory , Series on Ad- vances in Mathematics for Applied Sciences, World Scientific, Singapore, 1997

    1997

  15. [23]

    Monard, Efficient tensor tomography in fan-beam coordinates , Inverse Probl

    F. Monard, Efficient tensor tomography in fan-beam coordinates , Inverse Probl. Imaging, 10 (2016), pp. 433–459

    2016

  16. [24]

    Monard, Efficient tensor tomography in fan-beam coordinates

    F. Monard, Efficient tensor tomography in fan-beam coordinates. II: Attenuated transforms, Inverse Probl. Imaging, 12 (2018), pp. 433–460

    2018

  17. [25]

    Monard and G

    F. Monard and G. Bal , Inverse source problems in transport via attenuated tensor tomog- raphy, https://arxiv.org/abs/arXiv:1908.06508v1

    1908 arXiv

  18. [26]

    Natterer, The mathematics of computerized tomography, B

    F. Natterer, The mathematics of computerized tomography, B. G. Teubner, Stuttgart; John Wiley & Sons, Ltd., Chichester, 1986

    1986

  19. [27]

    Natterer, Inversion of the attenuated Radon transform , Inverse Problems, 17 (2001), pp

    F. Natterer, Inversion of the attenuated Radon transform , Inverse Problems, 17 (2001), pp. 113–119

    2001

  20. [28]

    R. G. Novikov , Une formule d’inversion pour la transformation d’un rayonnement X att´ enu´ e, C. R. Acad. Sci. Paris S´ er. I Math., 332 (2001), pp. 1059–1063

    2001

  21. [29]

    Ntziachristos and R

    V. Ntziachristos and R. Weissleder , Experimental three-dimensional fluorescence recon- struction of diffuse media by use of a normalized Born approximation , Opt. Lett., 26 (2001), pp. 893–895. NUMERICAL RECONSTRUCTION OF RADIATIVE SOURCES 21

    2001

  22. [30]

    Ntziachristos, A

    V. Ntziachristos, A. G. Yodh, M. Schnall, and B. Chance, Concurrent MRI and diffuse optical tomography of breast after indocyanine green enhancement , Proc. Natl. Acad. Sci. USA, 97 (2000), pp. 2767–2772

    2000

  23. [31]

    Radon, ¨Uber die bestimmung von funktionen durch ihre integralwerte l¨ angs gewisser man- nigfaltigkeiten, Berichte S¨ achsische Akademie der Wissenschaften zu Leipzig, Math.-Phys

    J. Radon, ¨Uber die bestimmung von funktionen durch ihre integralwerte l¨ angs gewisser man- nigfaltigkeiten, Berichte S¨ achsische Akademie der Wissenschaften zu Leipzig, Math.-Phys. Kl., 69 (1917), pp. 262–277. (translated : On the determination of functions from their integ...

    1917

  24. [32]

    Sadiq and A

    K. Sadiq and A. Tamasan, On the range characterization of the two-dimensional attenuated Doppler transform, SIAM J. Math. Anal., 47 (2015), pp. 2001–2021

    2015

  25. [33]

    Sadiq and A

    K. Sadiq and A. Tamasan , On the range of the attenuated Radon transform in strictly convex sets, Trans. Amer. Math. Soc., 367 (2015), pp. 5375–5398

    2015

  26. [34]

    C. E. Siewert, An inverse source problem in radiative transfer, J. Quant. Spectrosc. Radiat. Transfer, 50 (1993), pp. 603–609

    1993

  27. [35]

    Stefanov and G

    P. Stefanov and G. Uhlmann , An inverse source problem in optical molecular imaging , Anal. PDE, 1 (2008), pp. 115–126

    2008

  28. [36]

    P. A. Toft , The Radon Transform — Theory and Implementation , PhD thesis, Technical University of Denmark, 1996. Graduate School of Informatics, Kyoto University, Yoshida Honmachi, Sakyo-ku, Kyoto 606-8501, Japan E-mail address: fujiwara@acs.i.kyoto-u.ac.jp Johann Radon Insti...

    1996

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.