Pith. sign in

REVIEW 3 major objections 3 minor 39 references

Iterated graph Laplacian for image restoration problems

T0 review · 3 major / 3 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read The paper proves that rebuilding the graph Laplacian from each new reconstruction yields a convergent regularization loop: as noise tends to zero, all three iterated schemes approach the true solution set.

desk verdict Solid convergence theory for three iterated graph-Laplacian schemes, but the sparse-angle CT showcase runs under a hypothesis the paper does not verify — the authors admit it in Appendix A. read the letter →

arxiv 2607.17313 v1 pith:IJSHGTAC submitted 2026-07-19 math.NA cs.NA

Stefano Aleotti , Davide Bianchi , Florian Bossmann , Marco Donatelli , Pietro Maurino This is my paper · ORCID
classification math.NAcs.NA MSC 65F1065F2265J2068U1094A08
keywords graphLaplacianregularizationiteratedTikhonovimagerestorationinverseproblemsconvergenceanalysisgraph-minimizingsolutionscomputedtomographydeblurring
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 studies a feedback loop for image restoration: solve a Tikhonov-type problem whose regularizer is a graph Laplacian built from a reference image, then use the reconstruction as the new reference and repeat. It introduces three ways to run this loop—standard (the reference is the latest reconstruction), error-equation (the reference is an estimate of the reconstruction error), and mixed (a few standard steps followed by error-equation steps)—and proves that all three are convergent regularization methods. As the noise level δ goes to zero, the stopped iterates approach the set of exact solutions, with no uniqueness assumption and no source condition. The convergence rests on a joint coercivity condition between the forward operator and the graph Laplacian that must hold uniformly in the number of pixels and in the reference image, plus, for the two error-based schemes, a boundedness condition that holds automatically for grayscale images. The numerical experiments show that iterating sharpens fine details beyond a single graph-Laplacian step.

What carries the argument

The central object is the graph Laplacian Δ_z built from a reference image z via local Gaussian weights on a fixed spatial radius. The regularizer R^q_z(x) = (1/q)||Δ_z x||_q^q is updated by replacing z with the current reconstruction (or, in the error-based schemes, with an estimate of the error), so the prior sharpens as the iterate improves. The proof is carried by two ingredients: a uniform joint coercivity condition (Hypothesis 3.2) coupling K and Δ_z with constants independent of N and z, and a Lipschitz estimate for R^q_z with respect to the reference image (Corollary 2.3), which together give dimension-free a priori bounds and controlled passage to limits. The natural limit object is

What would settle it

For the sparse-angle CT operator used in Section 7.3, compute the optimal coercivity constant C_J(N) in (13) at increasing discretizations (e.g., 64, 128, 256, 512 pixels per side). If C_J(N) diverges as N grows, the dimension-free convergence theorems do not apply to that operator; a direct test would then run the standard and mixed schemes on noiseless simulated data with the prescribed stopping rule and check whether the stopped iterates actually approach S_gm(y) or S(y).

Watch

Extended reading notes

Core claim

The paper's central claim is that the iterated graph Laplacian is not a heuristic but a class of regularization methods with convergence guarantees in the vanishing-noise regime. For the standard scheme, Theorem 5.2 proves dist2(x^δ_{n(δ)}, S_gm(y)) → 0 as δ→0 under the a priori rule α_{n(δ)}→0 and δ²/α_{n(δ)}→0, where S_gm(y) is the set of graph-minimizing solutions: minimizers of the graph regularizer among exact solutions, for some reference image that is itself an exact solution. For the error-equation and mixed schemes, Theorems 6.4 and 6.6 prove dist2(x^δ_{m(δ)}, S(y)) → 0, provided the predecessor of the stopped iterate is uniformly bounded (which the paper notes is automatic under a

Load-bearing premise

The load-bearing premise is that the forward operator and the graph Laplacian together satisfy the joint coercivity bound ||u|| ≤ C_J(||Ku||_Y + ||Δ_z u||_q) with a constant C_J independent of the number of pixels and of the reference image; the paper verifies this uniformity only for a periodic blur model and a full-angle CT model, while the sparse-angle CT operator used in the experiments is explicitly left for a 'considerably more involved' analysis.

Editorial extensions

If this is right

  • With the a priori stopping rule α_n→0 and δ²/α_n→0, the standard scheme's stopped reconstructions are guaranteed to approach the graph-minimizing solution set; when K is injective this is ordinary convergence to the true image.
  • The error-equation and mixed schemes converge to the exact-solution set without uniqueness or source conditions, assuming only a uniform bound on the iterate preceding the stopping point — a bound the paper shows holds for grayscale images in [0,1].
  • The convergence constants are dimension-free, so the guarantees survive mesh refinement and do not degrade as images are discretized more finely.
  • For the mixed scheme, the standard phase does not need to have converged; it only supplies a bounded starting point and a residual for the error-equation phase.
  • The theorems give a theoretical foundation for iterated graph-Laplacian refinement from any initial reconstruction: the initial guess enters only through boundedness, not through a closeness or accuracy assumption.

Reading between the lines

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

  • The most direct open thread is the sparse-angle CT operator used in the experiments: Appendix A verifies uniform joint coercivity only for periodic blur and full-angle CT and explicitly excludes the sparse-angle case. If the optimal coercivity constant grows with resolution, the stated theorems would not cover the paper's own CT experiment, which would then be a finite-N heuristic.
  • The set-valued limit S_gm(y) invites a selection rule — for instance, choosing among exact solutions the one minimizing a secondary criterion — which would convert set convergence into pointwise convergence and give a principled way to handle non-uniqueness.
  • Because the error-equation scheme regularizes only the correction, its analysis should extend to other convex regularizers (e.g., total variation or learned data-driven regularizers) whenever the same uniform coercivity and Lipschitz-with-respect-to-reference properties hold.
  • The dimension-free coercivity condition could be tested empirically by computing, for increasing N, the ratio of ||u|| to ||Ku||+||Δ_z u|| over a basis of differences; if the ratio grows, the missing sparse-angle verification would show up in practice.
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

3 major / 3 minor

Summary. The paper proposes three iterative graph-Laplacian regularization schemes for linear image restoration: a standard scheme (Algorithm 1), an error-equation scheme (Algorithm 2), and a mixed scheme (Algorithm 3). In each, the graph Laplacian is constructed from a reference image that is updated as the iteration proceeds. The main theoretical contribution is a set of convergence theorems (Theorems 5.2, 6.4, and 6.6) showing that, under a uniform joint coercivity hypothesis (Hypothesis 3.2) and appropriate a priori stopping rules, the stopped reconstructions converge to the exact solution set (or, for the standard scheme, to the set of graph-minimizing solutions) as the noise level tends to zero. Numerical experiments on deblurring and computed tomography illustrate gains in reconstruction quality.

Significance. If the results hold, they provide a rigorous regularization-theoretic foundation for adaptively updating a graph-based regularizer, going beyond the heuristic use of iterated graph Laplacians in earlier work. The proofs are standard but carefully executed, with explicit constants and a sensible set-valued limit object S_gm(y). The paper also clearly identifies the additional uniform boundedness assumption needed for the error-equation and mixed schemes. However, the verification of the key coercivity hypothesis is incomplete for the operator used in the CT experiment, and the claim that the additional boundedness assumption holds automatically for grayscale images is not justified. These gaps affect the advertised scope of the theory and need to be addressed before the paper can be accepted in its present form.

major comments (3)
  1. [§7.3, Appendix A, Hypothesis 3.2] The convergence theorems (Theorems 5.2, 6.4, 6.6) are stated under Hypothesis 3.2, which requires a joint coercivity constant C_J independent of N. Appendix A verifies Hypothesis 3.2 only for a periodic blur model (q=2, Proposition A.2) and a full-angle CT model (Proposition A.3). The CT experiment in §7.3 uses a sparse-angle operator (60 angles), for which the paper provides only a fixed-N coercivity (Remark A.1) and states that the uniform verification is 'considerably more involved and lies outside the scope of this presentation.' This is an acknowledged gap: as written, the stated theorem does not directly apply to the showcased CT operator. The experiment can only be justified by a fixed-N analogue that is not stated in the theorems. Please state the fixed-N version (with C_{J,N}) explicitly and either prove the dimension-free bound for sparse-angle CT or temper the claim that the C
  2. [§6.3–6.4, Eq. (30)/(35), Remark 6.5, Algorithm 2] Theorems 6.4 and 6.6 depend on the uniform bound (30)/(35). The paper asserts in the Introduction and in Remark 6.5 that this bound 'holds automatically under the grayscale box constraint 0 ≤ x ≤ 1.' However, Algorithm 2 minimizes over the unconstrained space X; no projection onto the box is performed, and no proof is supplied that the unconstrained iterates satisfy such a bound uniformly in δ and the stopping index. This is load-bearing: in the proof of Theorem 6.4, boundedness of x^δ_m is obtained from the bound on h^δ_m together with (30); without (30), the cumulative reconstruction can drift. The authors should either add a box-constraint projection to Algorithms 2 and 3 and prove (30), or state (30)/(35) as an additional substantive hypothesis rather than an automatic consequence of grayscale images.
  3. [Theorems 5.2, 6.4, 6.6; §3] The theorems are formulated with constants independent of the discretization level N, and their proofs use 'finite-dimensional compactness' to extract convergent subsequences. This is valid only when N is fixed while δ→0. If the intended claim is a uniform-in-N statement (e.g., a joint limit δ→0 and N→∞), the compactness argument does not apply because the space X changes with N. Conversely, if N is fixed, the N-independence of C_J is not needed for the convergence conclusion itself. Please clarify the quantification over N in Theorems 5.2, 6.4, and 6.6 and adjust the wording in the abstract and introduction that suggests a dimension-free convergence result.
minor comments (3)
  1. [Throughout] There are many internal cross-reference errors: 'Theorem 3.2' should be 'Hypothesis 3.2' in several places (e.g., §3 text and Proposition 3.6 statement); 'Theorem 3.4' should be 'Definition 3.4' in §5.2; 'Theorem 2.3' in the proof of Corollary 3.7 should be 'Corollary 2.3'; 'Theorem 6.2' in the proof of Theorem 6.3 should be 'Proposition 6.2'; 'Theorem 3.5' in the proof of Theorem 5.2 should be 'Remark 3.5'; and Section 7 refers to 'Theorem A.1/A.2/A.3' but should refer to 'Remark A.1' and 'Propositions A.2/A.3'. Please correct throughout.
  2. [§7.3, Table 3] In Table 3, the GMSD value at x^δ_{nstd} (0.0587) is worse than at x^δ_1 (0.0523), although the text states that the standard iterations improve quality. This is not necessarily an error, but the authors should acknowledge or explain this metric behavior, since GMSD is discussed as a key indicator of detail recovery.
  3. [§7, first paragraph] The sentence 'For every fixed deblurring or CT matrix used below, Theorem A.1 gives the joint coercivity needed for the q=1 numerical problem, although its constant may depend on N' should be linked to the statement about fixed-N convergence; as written, it points to a remark, not a theorem, and the connection to the convergence theorems of Sections 5–6 is left implicit.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the convergence theorems are self-contained under stated hypotheses; the sparse-angle CT verification gap is a limitation, not a circular step.

full rationale

The paper's derivation chain is not circular. The central convergence claims (Theorems 5.2, 6.4, 6.6) are genuine mathematical implications from Hypotheses 3.1 and 3.2, with a priori parameter and stopping rules. The limit sets S_gm(y) and S(y) are defined independently of the iteration: S_gm(y) is the union over z in S(y) of R_z-minimizers over S(y), and the proofs identify cluster points as members of these sets using minimality, joint coercivity, and continuity of R_z. No fitted parameter is renamed as a prediction, and no theorem is imported from a self-citation as the load-bearing argument. Citations [5] and [11] provide background and prior algorithms, but the new convergence analysis is proved from the stated assumptions. The only notable weakness is that Hypothesis 3.2 is verified dimension-free only for a periodic blur model (q=2) and a full-angle CT model, while the numerical CT experiment in Section 7.3 uses a sparse-angle operator; Appendix A explicitly acknowledges that the uniform verification for that operator is 'considerably more involved and lies outside the scope of this presentation' and provides only fixed-N coercivity in Remark A.1. This is an applicability gap, not circularity, because the paper does not claim the sparse-angle experiment satisfies all hypotheses of the convergence theorems. Accordingly, the circularity score is 0.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new physical entities. Its free parameters are algorithm hyperparameters (R, σ, B_c, nstd) chosen by hand for the numerical tests; the central convergence theorems do not depend on their specific values as long as the graph weights are positive. The main load-bearing assumptions are the uniform joint coercivity (Hypothesis 3.2) and the additional boundedness condition for the error-based schemes, neither of which is proven for all experimental settings. The set S_gm(y) is a mathematical definition, not an invented entity.

free parameters (4)
  • Graph spatial radius R = R=5 (Tests 1, 2, CT), R=4 (Test 4)
    Chosen by hand; controls the spatial locality of the graph edges and, together with σ, the structure of the regularizer. No automated selection rule is given.
  • Gaussian parameter σ = σ=10^-3 (standard) and 10^-4 (error stage) in Tests 1/2; σ=5·10^-3 in CT; σ=2·10^-3 in NN test
    Chosen per test to set the sensitivity of edge weights to intensity differences; the reported reconstruction quality depends on it.
  • Contrast cap B_c = B_c > 1 (inactive) in the experiments
    Introduced in Eq. (5)/(6) to guarantee positive weights; chosen larger than 1 in all reported computations, so it does not affect the numerical results.
  • Number of standard iterations nstd = nstd=5 (Test 1), nstd=4 (Test 2), nstd=2 (CT, NN test)
    Selected at the empirical plateau of the standard scheme; affects the final mixed reconstruction quality.
assumptions (6)
  • domain assumption Hypothesis 3.1: existence of a uniformly bounded exact solution x_gt with ||x_gt||_2 ≤ M_gt independent of N
    Standard in regularization theory; the authors note it holds automatically for grayscale images in a fixed range. Invoked in the a priori estimates of Theorems 5.2, 6.4, 6.6.
  • domain assumption Hypothesis 3.2: uniform joint coercivity kuk_2 ≤ C_J(||Ku||_Y + ||Δ_z u||_q) with constants independent of N and z
    Morozov-type complementation condition; the load-bearing premise of the convergence theorems. Verified for a periodic blur model (q=2) and a full-angle CT model, but not for the sparse-angle CT used in the numerical experiments.
  • ad hoc to paper Uniform boundedness of the predecessor reconstruction (Eq. (30) and Eq. (35)): sup_{0<δ≤δ0} ||x^δ_{m(δ)-1}||_2 ≤ M_X
    Additional condition required by the error-equation and mixed schemes, not proven from the other hypotheses; the paper argues it holds for grayscale images but the algorithm does not explicitly enforce the box constraint.
  • domain assumption Assumption (41): the discretized full-angle CT operator satisfies kuk_2 ≤ C_R ||K_N u||_Y with C_R independent of N
    Used in Proposition A.3 to verify the dimension-free joint coercivity for full-angle CT; asserted as a convergence property of the discretization, not proven.
  • domain assumption For the periodic blur model: b̂_h(0) ≠ 0 and the graph contains the nearest-neighbor grid edges
    Assumed in Proposition A.2 to obtain the dimension-free coercivity estimate for the periodic convolution model.
  • standard math Standard finite-dimensional convex analysis and regularization theory (existence of minimizers, compactness of bounded sets, convexity of functionals)
    Involved throughout the proofs of Propositions 3.6, Corollary 3.7, and the convergence theorems.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Iterated graph Laplacian for image restoration problems." pith.science (2026). https://pith.science/paper/IJSHGTAC

@misc{pith2026260717313,
  author       = {Pith},
  title        = {Pith review of: Iterated graph Laplacian for image restoration problems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IJSHGTAC}},
  note         = {Machine review of arXiv:2607.17313}
}
read the original abstract

We study the graph Laplacian operator as a regularizer in a generalized Tikhonov framework for linear ill-posed problems. The Laplacian is updated iteratively from the current reconstruction, so that progressively sharper structural information about the solution is fed into the regularization term. We introduce three schemes: a standard one that rebuilds the Laplacian from each new iterate; an error-equation scheme that, following the error-based formulation of iterated Tikhonov regularization, builds the Laplacian from an estimate of the reconstruction error rather than of the image itself; and a mixed scheme combining the two. We establish convergence of all three schemes for noisy data under a priori parameter and stopping rules as the noise level tends to zero. Numerical experiments in two-dimensional computed tomography and image deblurring show consistent gains in reconstruction quality and sharper recovery of fine details.

Figures

Figures reproduced from arXiv: 2607.17313 by the authors.

Figure 1
Figure 1. Satellite deblurring test: (a) original image, (b) PSF with a [PITH_FULL_IMAGE:figures/full_fig_p018_1.png] view at source ↗
Figure 2
Figure 2. Restored images for Test 1: (a) graphLa+Tik (x δ 1 ), (b) standard iterative scheme (x δ nstd = x δ 5 ), (c) final reconstruction with mixed iterative scheme (x δ nstd+6). (a) (b) (c) (d) [PITH_FULL_IMAGE:figures/full_fig_p019_2.png] view at source ↗
Figure 3
Figure 3. Details of the restored images for Test 1: (a) detail of [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Restored images for Test 2: (a) graphLa+TV (x δ 1 ), (b) standard iterative scheme (x δ nstd = x δ 4 ), (c) final reconstruction with mixed iterative scheme (x δ nstd+2). (a) (b) (c) (d) [PITH_FULL_IMAGE:figures/full_fig_p020_4.png]
Figure 5
Figure 5. Figure 5: Details of the restored images for Test 2: (a) detail of [PITH_FULL_IMAGE:figures/full_fig_p020_5.png]
Figure 6
Figure 6. Figure 6: Test CT: (a) original image, (b) observed sinogram corrupted by noise. [PITH_FULL_IMAGE:figures/full_fig_p021_6.png]
Figure 7
Figure 7. Figure 7: Restored images for Test CT: (a) graphLa [PITH_FULL_IMAGE:figures/full_fig_p021_7.png]
Figure 8
Figure 8. Figure 8: Details of the restored images for Test CT: (a) detail of [PITH_FULL_IMAGE:figures/full_fig_p021_8.png]
Figure 9
Figure 9. Figure 9: Image deblurring and denoising with neural network initialization: (a) original image, (b) blurred and noisy [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 10
Figure 10. Figure 10: Zoomed details for image deblurring and denoising with neural network initialization: (a) original image, [PITH_FULL_IMAGE:figures/full_fig_p022_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 1 linked inside Pith

  1. [5]

    Improved impedance inversion by the iterated graph Laplacian

    Davide Bianchi, Florian Boßmann, Wenlong Wang, and Mingming Liu. Improved impedance inversion by the iterated graph Laplacian. arXiv preprint arXiv:2404.16324, 2024

    arXiv 2024

  2. [3]

    On the convergence of the iterative regularization method assisted by the graph Laplacian with early stopping

    Harshit Bajpai, Gaurav Mittal, and Ankik Kumar Giri. On the convergence of the iterative regularization method assisted by the graph Laplacian with early stopping. SIAM Journal on Imaging Sciences , 19(2):643–676, 2026

    2026

  3. [1]

    A nested primal–dual iterated Tikhonov method for regularized convex optimization

    Stefano Aleotti, Silvia Bonettini, Marco Donatelli, Marco Prato, and Simone Rebegoldi. A nested primal–dual iterated Tikhonov method for regularized convex optimization. Computational Optimization and Applications , 91(2):357–395, 2025

    2025

  4. [2]

    A variational framework for non-local image inpainting

    Pablo Arias, Vicent Caselles, and Guillermo Sapiro. A variational framework for non-local image inpainting. In Daniel Cremers, Y uri Boykov, Andrew Blake, and Frank R. Schmidt, editors, Energy Minimization Methods in Computer Vision and Pattern Recognition, volume 5681 of Lecture Notes in Computer Science , pages 345–358. Springer, Berlin, Heidelberg, 2009

    2009

  5. [4]

    Bauschke and Patrick L

    Heinz H. Bauschke and Patrick L. Combettes. Convex Analysis and Monotone Operator Theory in Hilbert Spaces. CMS Books in Mathematics. Springer, Cham, 2 edition, 2017

    2017

  6. [6]

    Graph Laplacian for image deblur- ring

    Davide Bianchi, Alessandro Buccini, Marco Donatelli, and Emma Randazzo. Graph Laplacian for image deblur- ring. Electronic Transactions on Numerical Analysis, 55:169–186, 2022

    2022

  7. [7]

    Iterated fractional Tikhonov regularization

    Davide Bianchi, Alessandro Buccini, Marco Donatelli, and Stefano Serra-Capizzano. Iterated fractional Tikhonov regularization. Inverse Problems, 31(5):055005, 2015

    2015

  8. [8]

    Graph approximation and generalized Tikhonov regularization for signal deblurring

    Davide Bianchi and Marco Donatelli. Graph approximation and generalized Tikhonov regularization for signal deblurring. In 2021 21st International Conference on Computational Science and Its Applications (ICCSA) , pages 93–100. IEEE, 2021

    2021

Show all 39 references
  1. [9]

    Convergence analysis and parameter estimation for the iterated arnoldi-tikhonov method

    Davide Bianchi, Marco Donatelli, Davide Furchí, and Lothar Reichel. Convergence analysis and parameter estimation for the iterated arnoldi-tikhonov method. Numerische Mathematik, 157(3):749–779, 2025

    2025

  2. [10]

    The iterated Golub–Kahan–Tikhonov method

    Davide Bianchi, Marco Donatelli, Davide Furchì, and Lothar Reichel. The iterated Golub–Kahan–Tikhonov method. BIT Numerical Mathematics, 66(1):18, 2026

    2026

  3. [11]

    A data-dependent regularization method based on the graph Laplacian

    Davide Bianchi, Davide Evangelista, Stefano Aleotti, Marco Donatelli, Elena Loli Piccolomini, and Wenbin Li. A data-dependent regularization method based on the graph Laplacian. SIAM Journal on Scientific Computing , 47(2):C369–C398, 2025

    2025

  4. [12]

    Uniformly convex neural networks and non-stationary iterated network Tikhonov (iNETT) method

    Davide Bianchi, Guanghao Lai, and Wenbin Li. Uniformly convex neural networks and non-stationary iterated network Tikhonov (iNETT) method. Inverse Problems, 39(5):055002, 2023

    2023

  5. [13]

    Graph Laplacian in ℓ2–ℓq regularization for image reconstruction

    Alessandro Buccini and Marco Donatelli. Graph Laplacian in ℓ2–ℓq regularization for image reconstruction. In 2021 21st International Conference on Computational Science and Its Applications (ICCSA) , pages 29–38. IEEE, 2021

    2021

  6. [14]

    Software for limited memory restarted ℓp–ℓq minimization methods using generalized Krylov subspaces

    Alessandro Buccini and Lothar Reichel. Software for limited memory restarted ℓp–ℓq minimization methods using generalized Krylov subspaces. Electronic Transactions on Numerical Analysis, 61:66–91, 2024

    2024

  7. [15]

    Chan and Hai-Xia Liang

    Raymond H. Chan and Hai-Xia Liang. Half-quadratic algorithm for ℓp–ℓq problems with applications to TV– ℓ1 image restoration and compressive sensing. In Andrés Bruhn, Thomas Pock, and Xue-Cheng Tai, editors, Efficient Algorithms for Global Optimization Methods in Computer Visio...

    2014

  8. [16]

    Square smoothing regularization matrices with accurate boundary condi- tions

    Marco Donatelli and Lothar Reichel. Square smoothing regularization matrices with accurate boundary condi- tions. Journal of Computational and Applied Mathematics , 272:334–349, 2014

    2014

  9. [17]

    Deep Wiener deconvolution: Wiener meets deep learning for image deblurring

    Jiangxin Dong, Stefan Roth, and Bernt Schiele. Deep Wiener deconvolution: Wiener meets deep learning for image deblurring. In Advances in Neural Information Processing Systems , volume 33, pages 1048–1059. Curran Associates, 2020

    2020

  10. [18]

    Regularization of Inverse Problems, volume 375 of Mathematics and its Applications

    Heinz Werner Engl, Martin Hanke, and Andreas Neubauer. Regularization of Inverse Problems, volume 375 of Mathematics and its Applications . Kluwer Academic Publishers, Dordrecht, 1996

    1996

  11. [19]

    Nonlocal linear image regularization and supervised segmentation

    Guy Gilboa and Stanley Osher. Nonlocal linear image regularization and supervised segmentation. Multiscale Modeling & Simulation, 6(2):595–630, 2007

    2007

  12. [20]

    Nonlocal operators with applications to image processing

    Guy Gilboa and Stanley Osher. Nonlocal operators with applications to image processing. Multiscale Modeling & Simulation, 7(3):1005–1028, 2009. 26 Iterated graph Laplacian for image restoration problems A P REPRINT

    2009

  13. [21]

    Rank-Deficient and Discrete Ill-Posed Problems: Numerical Aspects of Linear Inversion

    Per Christian Hansen. Rank-Deficient and Discrete Ill-Posed Problems: Numerical Aspects of Linear Inversion . Society for Industrial and Applied Mathematics, Philadelphia, PA, 1998

    1998

  14. [22]

    AIR Tools II: Algebraic iterative reconstruction methods, improved implementation

    Per Christian Hansen and Jakob Sauer Jørgensen. AIR Tools II: Algebraic iterative reconstruction methods, improved implementation. Numerical Algorithms, 79(1):107–137, 2018

    2018

  15. [23]

    Nagy, and Dianne P

    Per Christian Hansen, James G. Nagy, and Dianne P . O’Leary. Deblurring Images: Matrices, Spectra, and Filtering. Society for Industrial and Applied Mathematics, Philadelphia, PA, 2006

    2006

  16. [24]

    Majorization- minimization generalized Krylov subspace methods for ℓp–ℓq optimization applied to image restoration

    Guangxin Huang, Alessandro Lanza, Serena Morigi, Lothar Reichel, and Fiorella Sgallari. Majorization- minimization generalized Krylov subspace methods for ℓp–ℓq optimization applied to image restoration. BIT Numerical Mathematics, 57(2):351–378, 2017

    2017

  17. [25]

    Jie Huang, Marco Donatelli, and Raymond H. Chan. Nonstationary iterated thresholding algorithms for image deblurring. Inverse Problems and Imaging, 7(3):717–736, 2013

    2013

  18. [26]

    Nonstationary iterated Tikhonov regularization for ill-posed problems in Banach spaces

    Qinian Jin and Linda Stals. Nonstationary iterated Tikhonov regularization for ill-posed problems in Banach spaces. Inverse Problems, 28(10):104011, 2012

    2012

  19. [27]

    Nonstationary iterated Tikhonov regularization in Banach spaces with uniformly convex penalty terms

    Qinian Jin and Min Zhong. Nonstationary iterated Tikhonov regularization in Banach spaces with uniformly convex penalty terms. Numerische Mathematik, 127(3):485–513, 2014

    2014

  20. [28]

    Wojciechowski

    Matthias Keller, Daniel Lenz, and Radosław K. Wojciechowski. Graphs and Discrete Dirichlet Spaces , volume 358 of Grundlehren der mathematischen Wissenschaften. Springer, Cham, 2021

    2021

  21. [29]

    A generalized Krylov subspace method for ℓp–ℓq minimization

    Alessandro Lanza, Serena Morigi, Lothar Reichel, and Fiorella Sgallari. A generalized Krylov subspace method for ℓp–ℓq minimization. SIAM Journal on Scientific Computing , 37(5):S30–S50, 2015

    2015

  22. [30]

    Bertozzi

    Yifei Lou, Xiaoqun Zhang, Stanley Osher, and Andrea L. Bertozzi. Image recovery via nonlocal operators. Journal of Scientific Computing , 42(2):185–197, 2010

    2010

  23. [31]

    Moen, Baiyu Chen, David R

    Taylor R. Moen, Baiyu Chen, David R. Holmes, III, Xinhui Duan, Zhicong Y u, Lifeng Y u, Shuai Leng, Joel G. Fletcher, and Cynthia H. McCollough. Low-dose CT image and projection dataset. Medical Physics, 48(2):902– 911, 2021

    2021

  24. [32]

    Methods for Solving Incorrectly Posed Problems

    Vladimir Alekseevich Morozov. Methods for Solving Incorrectly Posed Problems. Springer, New Y ork, 1984

    1984

  25. [33]

    Nagy and Lucas Onisk

    James G. Nagy and Lucas Onisk. Mixed precision iterative refinement for linear inverse problems. SIAM Journal on Matrix Analysis and Applications , 47(1):388–411, 2026

    2026

  26. [34]

    The Mathematics of Computerized Tomography, volume 32 of Classics in Applied Mathematics

    Frank Natterer. The Mathematics of Computerized Tomography, volume 32 of Classics in Applied Mathematics . Society for Industrial and Applied Mathematics, Philadelphia, PA, 2001

    2001

  27. [35]

    Gabriel Peyré, Sébastien Bougleux, and Laurent D. Cohen. Non-local regularization of inverse problems. In David Forsyth, Philip Torr, and Andrew Zisserman, editors, Computer Vision – ECCV 2008 , pages 57–68. Springer, Berlin, Heidelberg, 2008

    2008

  28. [36]

    Rudin, Stanley Osher, and Emad Fatemi

    Leonid I. Rudin, Stanley Osher, and Emad Fatemi. Nonlinear total variation based noise removal algorithms. Physica D: Nonlinear Phenomena , 60(1–4):259–268, 1992

    1992

  29. [37]

    Springer, New Y ork, 2009

    Otmar Scherzer, Markus Grasmair, Harald Grossauer, Markus Haltmeier, and Frank Lenzen.V ariational Methods in Imaging, volume 167 of Applied Mathematical Sciences. Springer, New Y ork, 2009

    2009

  30. [38]

    Wufeng Xue, Lei Zhang, Xuanqin Mou, and Alan C. Bovik. Gradient magnitude similarity deviation: A highly efficient perceptual image quality index. IEEE Transactions on Image Processing, 23(2):684–695, 2014

    2014

  31. [39]

    Bregmanized nonlocal regularization for deconvolution and sparse reconstruction

    Xiaoqun Zhang, Martin Burger, Xavier Bresson, and Stanley Osher. Bregmanized nonlocal regularization for deconvolution and sparse reconstruction. SIAM Journal on Imaging Sciences , 3(3):253–276, 2010. 27

    2010

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.