Pith. sign in

REVIEW 3 major objections 3 minor 67 references

An incremental algorithm for non-convex AI-enhanced medical image processing

T0 review · 3 major / 3 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This paper claims that alternating a pretrained neural network's guess with a few model-based iterations solves non-convex TpV medical imaging problems faster and more stably than either approach alone.

desk verdict A practical hybrid that delivers real speedups and strong deblurring, but the abstract overclaims accuracy and theoretical guarantees that the paper's own convergence section disclaims. read the letter →

arxiv 2505.08324 v1 pith:BFYOL5XO submitted 2025-05-13 cs.CV cs.NAmath.NA

Elena Morotti This is my paper · ORCID
classification cs.CVcs.NAmath.NA MSC 68U1065K1068T0790C26
keywords non-convexoptimizationtotalp-variationincrementalreweightedalgorithmdeeplearninginitializationmedicalimagedeblurringsparse-viewCTground-truth-freetraininghybridmodel-basedreconstruction
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

Medical image reconstruction from blurred or subsampled measurements is posed here as a non-convex inverse problem whose regularizer, the total $p$-variation ($\mathrm{TpV}$) norm of the gradient image, approximates the $\ell_0$ quasi-norm and therefore has many local minima. The paper claims that incDG, a hybrid algorithm alternating a pretrained network's refined guess with a few model-based iterations per stage, efficiently approximates the $\ell_0$-optimal solution of this problem. The network provides a strong initialization, while the iterative solver keeps the reconstruction consistent with the measured data and stabilizes the output. In deblurring and sparse-view CT experiments, incDG reports lower relative error and higher SSIM than a purely iterative incremental solver and a purely network-based variant, and it keeps most of that quality when trained without ground truth. If true, this gives medical imaging a way to combine the speed of deep learning with the reliability of model-based optimization.

What carries the argument

The central object is the incremental homotopy schedule of subproblems, where the rule $p^{(h+1)} = p^{(h)}\alpha_p$ pushes the prior from convex total variation ($p=1$) toward the $\ell_0$ quasi-norm, and $\lambda^{(h+1)} = \lambda^{(h)} f^{(h)}/f^{(h-1)}$ adapts the regularization weight from the objective's successive values. Each stage is solved by an iterative reweighted $\ell_1$ majorization of the non-convex $\mathrm{TpV}$ term, with weights $w_i = p/(|Dx|_i^{1-p}+\xi)$, using only five Chambolle-Pock iterations; a ResUNet (a U-shaped convolutional network with residual skip connections) maps the previous iterate to a refined guess before those iterations. The mechanism works because the network lowers the cost of each stage, the majorization keeps each subproblem convex, and the schedule gradually removes regularization while sharpening the sparsity prior.

What would settle it

Take the incDG test sets, replace the network output at every stage with either the raw current iterate or a fixed blurred image, and measure RE/SSIM: if the gap to incTpV vanishes, the claimed advantage is due to the neural initialization rather than the incremental mechanism. Alternatively, perturb the scheduler parameters $\alpha_p$ and the iteration budget $K$ slightly and check whether final RE/SSIM jumps discontinuously, which would indicate the result is a tuned artifact of the schedule.

Watch

Extended reading notes

Core claim

On its own terms, the paper establishes a constructive claim: the $\ell_0$-optimal solution of a $\mathrm{TpV}$-regularized imaging problem can be approximated by running a sequence of cheap convex subproblems, each initialized by a convolutional network and refined by five Chambolle-Pock iterations, while the sparsity parameter $p$ is driven geometrically from 1 toward 0 and the regularization weight $\lambda$ is adapted from the ratio of successive objective values. The resulting incDG algorithm is reported to reach deblurring and CT reconstructions with better RE and SSIM than the purely model-based incTpV and the purely learned incNN, and with far fewer iterations than incTpV (20-30 versus 270-3100). The paper further claims that the model-based refinement is what confers stability: incNN's SSIM fluctuates widely and it hallucinates streaking artifacts, whereas incDG's metric spread stays narrow. Finally, training the networks against incTpV solutions instead of ground truth degrades quality only slightly, which the paper offers as evidence that the method can be deployed where clean references do not exist.

Load-bearing premise

The load-bearing premise is that a handful of inexact inner iterations (five per stage), guided by the network's guess and the adaptive schedule, lands in a good local minimum of the non-convex $\mathrm{TpV}$ objective; the paper explicitly notes that convergence of this inexact incremental algorithm is not proven.

Editorial extensions

If this is right

  • A model-based solver can run in a clinical time budget without losing accuracy: incDG finishes CT reconstruction in about 4.5 seconds versus 53 seconds for the incremental model-based solver, with comparable or better image quality.
  • A short model-based refinement protects any learned reconstruction against the instability that plagues standalone networks: incNN's SSIM swings widely and produces hallucinated streaks, while incDG's metrics stay concentrated.
  • Ground-truth-free training is a viable route: using model-based incremental solutions as training targets yields reconstructions almost as good as supervised training, so the method can be applied when no clean anatomical images are available.
  • The incremental schedule itself improves on fixed-parameter $\mathrm{TpV}$ solutions, producing brighter, better-contrasted images that preserve low-contrast boundaries, which is the clinically relevant goal of enhancing subtle findings.

Reading between the lines

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

  • The same architecture could likely accelerate other non-convex sparsity priors (log-sum, atan, or $\ell_p$ with different $p$ paths), since the machinery only needs a convex majorizer and a network that maps current iterates toward the target; the paper demonstrates only $\mathrm{TpV}$.
  • Because most of incDG's runtime is fixed overhead (network forward pass and initialization), a lighter network or fused execution could push reconstruction toward real-time without changing the optimization logic.
  • The per-stage networks are trained independently with an MSE loss; a testable extension is to train all cascaded networks jointly with a loss on the final output, which could exploit the cascade structure and possibly improve end-to-end accuracy.
  • The stability claim implies a concrete diagnostic: under distribution shift or adversarial perturbation, incDG's worst-case SSIM should degrade gracefully while incNN's drops sharply; measuring that gap would validate whether the model-based iterations genuinely cap hallucination.
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 incDG, a hybrid algorithm for non-convex TpV-regularized inverse problems in medical imaging. It combines an incremental reweighted-ell_1 scheme with per-stage ResUNet 'Deep Guess' initializations followed by a few Chambolle-Pock iterations. The stated claims are that incDG efficiently approximates the ell_0-optimal solution, outperforms both conventional iterative solvers and deep-learning-based methods in accuracy and stability, and can be trained without ground truth. Experiments cover deblurring on COULE images with Brain CT generalization tests, and sparse-view CT reconstruction on the Mayo Clinic dataset.

Significance. If the main claims were fully established, incDG would be a practically valuable hybrid: it mitigates the instability of pure network outputs by adding a small number of model-based iterations, and the ground-truth-free training variant would ease clinical deployment. The paper has real strengths: it uses external public datasets, compares incNN, incTpV, and incDG under matched parameter settings, reports wall-clock timings, and provides a clear empirical demonstration that a few MB iterations stabilize a network-based reconstruction. However, the central theoretical claim is explicitly disclaimed in Section 2.4, and the CT results contradict the abstract's blanket accuracy claim. The paper needs either a convergence or suboptimality analysis for the inexact schedule, or a reframing of the contribution as heuristic acceleration with empirical claims restricted to speed and stability.

major comments (3)
  1. [§2.4, Algorithm 3] Section 2.4 states that 'the convergence of the resulting algorithm is not proven' and that in the authors' setting each TpV problem is stopped after only one iteration. Algorithm 3 then fixes kCP=5 and uses schedulers K=[5,5,5,5] for deblurring or [5,5,5,5,5,5] for CT. No error bound is given for this truncation, and no result shows that the final iterate is stationary for problem (15) or a local minimizer of the TpV functional. Consequently the abstract's 'efficiently approximates the ell_0-optimal solution' and Section 5's 'theoretical guarantees of model-based optimization' are unsupported. Please either provide a convergence or suboptimality analysis for the inexact schedule, or explicitly reframe incDG as a heuristic acceleration and limit the claims to the empirical results.
  2. [Figure 8, Figure 9, Abstract] The Abstract claims that 'incDG outperforms both conventional iterative solvers and deep learning-based methods, achieving superior accuracy and stability.' In the CT experiments, however, Figure 8's caption states that incTpV 'outperforms the fast methods, achieving the best RE and SSIM values on the displayed images,' and Figure 9's boxplots show incTpV with the best metric distributions over the test set. Thus the accuracy part of the claim is contradicted in the CT setting; incDG's advantage there is speed (Table 2) and stability relative to incNN, not accuracy relative to incTpV. Please restrict the accuracy claim to the deblurring experiments or revise the abstract and conclusions accordingly.
  3. [§4.1, Table 2] The performance of incDG depends on several hand-set hyperparameters (lambda_0, alpha_p, H, scheduler K, kCP) that are fixed to 'achieve good performance on the training samples,' with no sensitivity study. Because Section 2.4 provides no convergence guarantee, the observed gap between incDG and incTpV/incNN could be a tuned artifact of the scheduler and network cascade rather than an intrinsic property of the algorithm. Please report a sensitivity analysis for at least alpha_p and the scheduler K, and quantify how the results change with kCP, to substantiate the claimed robustness.
minor comments (3)
  1. [Algorithm 1, line 10] The stopping condition 'or count< kIR' is inverted: since count starts at 0, this condition is true immediately and would stop the loop at once. It should be 'or count>= kIR' (or an equivalent maximum-iteration check), with count accumulating kCP per outer iteration as described.
  2. [Equations (13), (14), §2.4] Equations (13) and (14) use '∀j' where the running index is h; moreover, Section 2.4 refers to 'Equation (5)' for the lambda and p updates, but the relevant formulation is Equation (15).
  3. [General typography] There are typographical slips such as 'tollerances' in Algorithms 2 and 3 and 'regularizaion' in Section 2.1 that should be corrected before publication.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: the core optimization is a standard reweighted-l1 majorization, and the performance claims are tested on external benchmarks; the acknowledged convergence gap is a correctness issue, not circularity.

full rationale

The derivation chain is self-contained rather than circular. The non-convex TpV objective is handled by the standard Iterative Reweighted majorization of Equations (9)-(11), with weights defined explicitly in Equation (12) from the current iterate and sparsity parameter; this is not defined in terms of the claimed output. The incremental schedules for lambda(h) and p(h) in Equations (13)-(14) are taken from prior work, but the paper re-derives the subproblem (15) for its own TpV regularizer and does not claim that the cited results prove convergence of its inexact version. Indeed, Section 2.4 explicitly states: 'Although the convergence of the resulting algorithm is not proven, it has demonstrated empirical efficiency' - this is an acknowledged missing proof, not a circular reduction. The Deep Guess component of [35] is a self-citation by the same research group, but the paper describes the network architecture, training losses, and experimental protocol in sufficient detail that the reported comparisons against incTpV, incNN, and standard TpV on COULE, Brain CT, and Mayo Clinic data are external empirical evidence rather than a restatement of the citation's conclusions. The ground-truth-free training targets in Equations (19)-(20) are incTpV solutions, but this does not force incDG to outperform incTpV; in fact, the CT results report that incTpV achieves the best RE and SSIM, so the central comparison is not manufactured. No equation is shown to be equivalent to its own input, no fitted parameter is renamed as a prediction, and no uniqueness theorem is imported to exclude alternatives. The main weakness is the unsupported convergence claim for the inexact incremental solver, which is a correctness risk but not circularity.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central method rests on three empirical assumptions: the incremental p/lambda schedule finds a good sparse basin, the deep guesses land in that basin, and five inner iterations per stage are sufficient. The mathematical majorization and convex subproblem structure are standard. No new physical or mathematical entities are introduced.

free parameters (6)
  • lambda_0 = 0.5 (deblurring), 0.01 (CT)
    Initial regularization weight; Section 4.1 states values were set to achieve good performance on training samples on average.
  • alpha_p = 0.5 (deblurring), 0.7 (CT)
    Decay factor for sparsity exponent p in Equation (14); tuned per task and directly controls the final effective p.
  • H = 4 (deblurring), 6 (CT)
    Number of incremental stages; chosen manually and affects the final p value after geometric decay.
  • scheduler K = deblurring [100,100,50,10] / [5,5,5,5]; CT [200,500,500,500,700,700] / [5,5,5,5,5,5]
    Per-stage Chambolle-Pock iteration budgets; central to both runtime and quality comparisons, with no sensitivity analysis reported.
  • xi = 2e-3
    Smoothing parameter in the reweighting weights of Equation (12); fixed by hand and used in all experiments.
  • kCP = 5
    Number of inner Chambolle-Pock iterations per weighted subproblem; fixed globally and critical to the runtime advantage.
assumptions (5)
  • standard math Kurdyka-Lojasiewicz property holds for the TpV objective, so the path of lambda-decreasing penalized solutions converges as shown in [33].
    Invoked in Section 2.4 to assert theoretical grounding, but the paper concedes the implemented inexact incremental variant has no convergence proof.
  • domain assumption The schedule p(h+1)=alpha_p p(h) and adaptive lambda(h) steers iterates toward a good ell_0-like local minimum.
    Empirical rule borrowed from [33,41]; no theorem links this schedule to the ell_0 optimum for the TpV objective.
  • domain assumption Network outputs used as initial guesses lie in the basin of attraction of a good local minimum after few Chambolle-Pock iterations.
    Core mechanism of the Deep Guess idea; only empirical validation is given, with no guarantee for the per-step insertion.
  • domain assumption Each weighted ell_1 subproblem can be approximated with kCP=5 primal-dual iterations and still preserve the incremental path.
    The runtime advantage depends on this budget; no error bound or sensitivity analysis is provided.
  • standard math The objective is semi-algebraic and definable in an o-minimal structure, making the KL theory applicable.
    Plausible by composition of semi-algebraic terms, but not verified in detail for the smoothed reweighted formulation used in the code.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An incremental algorithm for non-convex AI-enhanced medical image processing." pith.science (2026). https://pith.science/paper/BFYOL5XO

@misc{pith2026250508324,
  author       = {Pith},
  title        = {Pith review of: An incremental algorithm for non-convex AI-enhanced medical image processing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BFYOL5XO}},
  note         = {Machine review of arXiv:2505.08324}
}
abstract

Solving non-convex regularized inverse problems is challenging due to their complex optimization landscapes and multiple local minima. However, these models remain widely studied as they often yield high-quality, task-oriented solutions, particularly in medical imaging, where the goal is to enhance clinically relevant features rather than merely minimizing global error. We propose incDG, a hybrid framework that integrates deep learning with incremental model-based optimization to efficiently approximate the $\ell_0$-optimal solution of imaging inverse problems. Built on the Deep Guess strategy, incDG exploits a deep neural network to generate effective initializations for a non-convex variational solver, which refines the reconstruction through regularized incremental iterations. This design combines the efficiency of Artificial Intelligence (AI) tools with the theoretical guarantees of model-based optimization, ensuring robustness and stability. We validate incDG on TpV-regularized optimization tasks, demonstrating its effectiveness in medical image deblurring and tomographic reconstruction across diverse datasets, including synthetic images, brain CT slices, and chest-abdomen scans. Results show that incDG outperforms both conventional iterative solvers and deep learning-based methods, achieving superior accuracy and stability. Moreover, we confirm that training incDG without ground truth does not significantly degrade performance, making it a practical and powerful tool for solving non-convex inverse problems in imaging and beyond.

Figures

Figures reproduced from arXiv: 2505.08324 by the authors.

Figure 1
Figure 1. Sketch of the considered incremental approaches, applied to tomographic image reconstruction. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Plots of the p-norm functions ||x||p for different values of p. The curves correspond to different values of p: p = 0 (blue), p = 1 (brown), and intermediate values approaching p → 0. They illustrate the loss of convexity for p < 1 and the non-differentiability characteristic of the ℓ0. As anticipated, setting the prior function R(x) with the ℓ0 quasi-norm of the gradient image should be optimal to preserve sharp ed… view at source ↗
Figure 3
Figure 3. Scheme of the ResUNet architecture used to define the image-to-image operators [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Experiments performed on a COULE test sample for image deblurring and denoising. On the top, from left [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: Experiments performed on COULE test samples for image deblurring and denoising. At the top: one ground [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: Experiments performed on Brain CT test samples for image deblurring and denoising. At the top, from left [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Experiments performed on Mayo Clinic test samples for CT image reconstruction from subsampled data. [PITH_FULL_IMAGE:figures/full_fig_p016_7.png]
Figure 8
Figure 8. Figure 8: Experiments performed on Mayo Clinic test samples for CT image reconstruction from subsampled data. [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: Experiments performed on Mayo Clinic test samples for CT image reconstruction from subsampled data. [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

67 extracted references · 57 canonical work pages

  1. [1]

    The gap between theory and practice in function approximation with deep neural networks

    Ben Adcock and Nick Dexter. The gap between theory and practice in function approximation with deep neural networks. SIAM Journal on Mathematics of Data Science, 3(2):624–655, 2021

    work page 2021

  2. [2]

    Plug-and-play methods for magnetic resonance imaging: Using denoisers for image recovery

    Rizwan Ahmad, Charles A Bouman, Gregery T Buzzard, Stanley Chan, Sizhuo Liu, Edward T Reehorst, and Philip Schniter. Plug-and-play methods for magnetic resonance imaging: Using denoisers for image recovery. IEEE signal processing magazine, 37(1):105–116, 2020

    work page 2020

  3. [3]

    Solving inverse problems using data- driven models

    Simon Arridge, Peter Maass, Ozan ¨Oktem, and Carola-Bibiane Sch¨onlieb. Solving inverse problems using data- driven models. Acta Numerica, 28:1–174, 2019

    work page 2019

  4. [4]

    Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward–backward splitting, and regularized gauss–seidel methods

    Hedy Attouch, J ´erˆome Bolte, and Benar Fux Svaiter. Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward–backward splitting, and regularized gauss–seidel methods. Math- ematical programming, 137(1):91–129, 2013

    work page 2013

  5. [5]

    The mathematics of adversarial attacks in ai–why deep learning is unstable despite the existence of stable neural networks.arXiv preprint arXiv:2109.06098, 2021

    Alexander Bastounis, Anders C Hansen, and Verner Vla ˇci´c. The mathematics of adversarial attacks in ai–why deep learning is unstable despite the existence of stable neural networks.arXiv preprint arXiv:2109.06098, 2021

    arXiv 2021

  6. [6]

    Discrete radon transform

    Gregory Beylkin. Discrete radon transform. IEEE transactions on acoustics, speech, and signal processing , 35(2):162–172, 1987

    work page 1987

  7. [7]

    A new twist: Two-step iterative shrinkage/thresholding algo- rithms for image restoration

    Jos ´e M Bioucas-Dias and M ´ario AT Figueiredo. A new twist: Two-step iterative shrinkage/thresholding algo- rithms for image restoration. IEEE Transactions on Image processing, 16(12):2992–3004, 2007

    work page 2007

  8. [8]

    A scaled gradient projection method for constrained image deblurring

    Silvia Bonettini, Riccardo Zanella, and Luca Zanni. A scaled gradient projection method for constrained image deblurring. Inverse problems, 25(1):015002, 2008

    work page 2008

Show all 67 references
  1. [9]

    Iterated hard shrinkage for minimization problems with sparsity constraints

    Kristian Bredies and Dirk A Lorenz. Iterated hard shrinkage for minimization problems with sparsity constraints. SIAM Journal on Scientific Computing, 30(2):657–683, 2008

    2008

  2. [10]

    Bubba, Luca Calatroni, Ambra Catozzi, Serena Crisci, Thomas Pock, Monica Pragliola, Siiri Rautio, Danilo Riccio, and Andrea Sebastiani

    Tatiana A. Bubba, Luca Calatroni, Ambra Catozzi, Serena Crisci, Thomas Pock, Monica Pragliola, Siiri Rautio, Danilo Riccio, and Andrea Sebastiani. Bilevel learning of regularization models and their discretization for image deblurring and super-resolution. In Alessandro Benfen...

    2024

  3. [11]

    Learning the invisible: A hybrid deep learning-shearlet framework for limited angle computed tomography

    Tatiana A Bubba, Gitta Kutyniok, Matti Lassas, Maximilian M ¨arz, Wojciech Samek, Samuli Siltanen, and Vig- nesh Srinivasan. Learning the invisible: A hybrid deep learning-shearlet framework for limited angle computed tomography. Inverse Problems, 35(6):064002, 2019

    2019

  4. [12]

    Bilevel approaches for learning of variational imaging models.Variational Methods: In Imaging and Geometric Control, 18(252):2, 2017

    Luca Calatroni, Chung Cao, Juan Carlos De Los Reyes, Carola-Bibiane Sch¨onlieb, and Tuomo Valkonen. Bilevel approaches for learning of variational imaging models.Variational Methods: In Imaging and Geometric Control, 18(252):2, 2017

    2017

  5. [13]

    Enhancing sparsity by reweighted L1 minimization

    Emmanuel J Cand `es, Michael B Wakin, and Stephen P Boyd. Enhancing sparsity by reweighted L1 minimization. Journal of Fourier analysis and applications, 2008

    2008

  6. [14]

    Plug-and-play gradient- based denoisers applied to ct image enhancement

    Pasquale Cascarano, Elena Loli Piccolomini, Elena Morotti, and Andrea Sebastiani. Plug-and-play gradient- based denoisers applied to ct image enhancement. Applied Mathematics and Computation, 422:126967, 2022

    2022

  7. [15]

    Gpu acceleration of a model-based iterative method for digital breast tomosynthesis

    R Cavicchioli, J Cheng Hu, E Loli Piccolomini, E Morotti, and L Zanni. Gpu acceleration of a model-based iterative method for digital breast tomosynthesis. Scientific Reports, 2020

    2020

  8. [16]

    A first-order primal-dual algorithm for convex problems with applications to imaging

    Antonin Chambolle and Thomas Pock. A first-order primal-dual algorithm for convex problems with applications to imaging. Journal of mathematical imaging and vision, 40:120–145, 2011

    2011

  9. [17]

    Exact reconstruction of sparse signals via nonconvex minimization

    Rick Chartrand. Exact reconstruction of sparse signals via nonconvex minimization. IEEE Signal Processing Letters, 14(10):707–710, 2007. 19 A PREPRINT - MAY 14, 2025

    2007

  10. [18]

    Fast algorithms for nonconvex compressive sensing: Mri reconstruction from very few data

    Rick Chartrand. Fast algorithms for nonconvex compressive sensing: Mri reconstruction from very few data. In 2009 IEEE international symposium on biomedical imaging: from nano to macro, pages 262–265. IEEE, 2009

    2009

  11. [19]

    Iteratively reweighted least squares minimization for sparse recovery

    Ingrid Daubechies, Ronald DeV ore, Massimo Fornasier, and C Sinan G ¨unt¨urk. Iteratively reweighted least squares minimization for sparse recovery. Communications on Pure and Applied Mathematics: A Journal Issued by the Courant Institute of Mathematical Sciences, 63(1):1–38, 2010

    2010

  12. [20]

    To be or not to be stable, that is the question: understanding neural networks for inverse problems

    Davide Evangelista, Elena Loli Piccolomini, Elena Morotti, and James G Nagy. To be or not to be stable, that is the question: understanding neural networks for inverse problems. SIAM Journal on Scientific Computing , 47(1):C77–C99, 2025

    2025

  13. [21]

    RISING: A new framework for model- based few-view CT image reconstruction with deep learning

    Davide Evangelista, Elena Morotti, and Elena Loli Piccolomini. RISING: A new framework for model- based few-view CT image reconstruction with deep learning. Computerized Medical Imaging and Graphics , 103:102156, 2023

    2023

  14. [22]

    Ambiguity in solving imaging inverse problems with deep-learning-based operators

    Davide Evangelista, Elena Morotti, Elena Loli Piccolomini, and James Nagy. Ambiguity in solving imaging inverse problems with deep-learning-based operators. Journal of Imaging, 9(7):133, 2023

    2023

  15. [23]

    Gradient projection for sparse reconstruction: Application to compressed sensing and other inverse problems

    M ´ario AT Figueiredo, Robert D Nowak, and Stephen J Wright. Gradient projection for sparse reconstruction: Application to compressed sensing and other inverse problems. IEEE Journal of selected topics in signal pro- cessing, 1(4):586–597, 2007

    2007

  16. [24]

    Solving inverse problems with deep neural networks- robustness included

    Martin Genzel, Jan Macdonald, and Maximilian Marz. Solving inverse problems with deep neural networks- robustness included. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022

    2022

  17. [25]

    Deep residual learning for compressed sensing ct reconstruction via persistent homology analysis

    Yo Seob Han, Jaejun Yoo, and Jong Chul Ye. Deep residual learning for compressed sensing ct reconstruction via persistent homology analysis. arXiv preprint arXiv:1611.06391, 2016

    2016 arXiv

  18. [26]

    Framing u-net via deep convolutional framelets: Application to sparse-view ct

    Yoseob Han and Jong Chul Ye. Framing u-net via deep convolutional framelets: Application to sparse-view ct. IEEE transactions on medical imaging, 37(6):1418–1429, 2018

    2018

  19. [27]

    Deep learning-based solvability of underdetermined inverse problems in medical imaging.Medical Image Analysis, 69:101967, 2021

    Chang Min Hyun, Seong Hyeon Baek, Mingyu Lee, Sung Min Lee, and Jin Keun Seo. Deep learning-based solvability of underdetermined inverse problems in medical imaging.Medical Image Analysis, 69:101967, 2021

    2021

  20. [28]

    Resunet++: An advanced architecture for medical image segmentation

    Debesh Jha, Pia H Smedsrud, Michael A Riegler, Dag Johansen, Thomas De Lange, P˚al Halvorsen, and H˚avard D Johansen. Resunet++: An advanced architecture for medical image segmentation. In 2019 IEEE international symposium on multimedia (ISM), pages 225–2255. IEEE, 2019

    2019

  21. [29]

    Plug-and-play methods for integrating physical and learned models in computational imaging: Theory, algorithms, and applications

    Ulugbek S Kamilov, Charles A Bouman, Gregery T Buzzard, and Brendt Wohlberg. Plug-and-play methods for integrating physical and learned models in computational imaging: Theory, algorithms, and applications. IEEE Signal Processing Magazine, 40(1):85–97, 2023

    2023

  22. [30]

    Learning regularization parameter-maps for variational image reconstruction using deep neural networks and algorithm unrolling

    Andreas Kofler, Fabian Altekr ¨uger, Fatima Antarou Ba, Christoph Kolbitsch, Evangelos Papoutsellis, David Schote, Clemens Sirotenko, Felix Frederik Zimmermann, and Kostas Papafitsoros. Learning regularization parameter-maps for variational image reconstruction using deep neur...

    2023

  23. [31]

    Brain tumor segmentation of the flair mri images using novel resunet

    P Santosh Kumar, VP Sakthivel, Manda Raju, and PD Sathya. Brain tumor segmentation of the flair mri images using novel resunet. Biomedical Signal Processing and Control, 82:104586, 2023

    2023

  24. [32]

    On gradients of functions definable in o-minimal structures

    Krzysztof Kurdyka. On gradients of functions definable in o-minimal structures. In Annales de l’institut Fourier, volume 48, pages 769–783, 1998

    1998

  25. [33]

    A nonconvex penalization algorithm with automatic choice of the regularization parameter in sparse imaging

    D Lazzaro, E Loli Piccolomini, and F Zama. A nonconvex penalization algorithm with automatic choice of the regularization parameter in sparse imaging. Inverse Problems, 35(8):084002, 2019

    2019

  26. [34]

    A fast total variation-based iterative algorithm for digital breast tomosynthesis image reconstruction

    E Loli Piccolomini and et al. A fast total variation-based iterative algorithm for digital breast tomosynthesis image reconstruction. Journal of Algorithms & Computational Technology, 2016

    2016

  27. [35]

    Deep guess acceleration for explainable image reconstruction in sparse-view ct

    Elena Loli Piccolomini, Davide Evangelista, and Elena Morotti. Deep guess acceleration for explainable image reconstruction in sparse-view ct. Computerized Medical Imaging and Graphics, page 102530, 2025

    2025

  28. [36]

    A model-based optimization framework for iterative digital breast tomosynthesis image reconstruction

    Elena Loli Piccolomini and Elena Morotti. A model-based optimization framework for iterative digital breast tomosynthesis image reconstruction. Journal of Imaging, 2021

    2021

  29. [37]

    Tu-fg-207a-04: Overview of the Low Dose CT Grand Challenge

    C McCollough. Tu-fg-207a-04: Overview of the Low Dose CT Grand Challenge. Medical physics , 43(6Part35):3759–3760, 2016

    2016

  30. [38]

    Sparse recovery using smoothed\ell 0 (SL0): Convergence analysis

    Hosein Mohimani, Massoud Babaie-Zadeh, Irina Gorodnitsky, and Christian Jutten. Sparse recovery using smoothed\ell 0 (SL0): Convergence analysis. arXiv preprint arXiv:1001.5073, 2010

    2010 arXiv

  31. [39]

    A fast approach for overcomplete sparse de- composition based on smoothed\ell 0 norm

    Hosein Mohimani, Massoud Babaie-Zadeh, and Christian Jutten. A fast approach for overcomplete sparse de- composition based on smoothed\ell 0 norm. IEEE Transactions on Signal Processing, 57(1):289–301, 2008. 20 A PREPRINT - MAY 14, 2025

    2008

  32. [40]

    Algorithm unrolling: Interpretable, efficient deep learning for signal and image processing

    Vishal Monga, Yuelong Li, and Yonina C Eldar. Algorithm unrolling: Interpretable, efficient deep learning for signal and image processing. IEEE Signal Processing Magazine, 38(2):18–44, 2021

    2021

  33. [41]

    An iterative l{1}-based image restoration algorithm with an adap- tive parameter estimation

    Laura B Montefusco and Damiana Lazzaro. An iterative l{1}-based image restoration algorithm with an adap- tive parameter estimation. IEEE Transactions on Image Processing, 21(4):1676–1686, 2011

    2011

  34. [42]

    Fast sparse image reconstruction using adaptive nonlinear filtering

    Laura B Montefusco, Damiana Lazzaro, and Serena Papi. Fast sparse image reconstruction using adaptive nonlinear filtering. IEEE Transactions on Image Processing, 20(2):534–544, 2010

    2010

  35. [43]

    A fast algorithm for nonconvex approaches to sparse recovery problems

    Laura B Montefusco, Damiana Lazzaro, and Serena Papi. A fast algorithm for nonconvex approaches to sparse recovery problems. Signal Processing, 93(9):2636–2647, 2013

    2013

  36. [44]

    A green prospective for learned post-processing in sparse-view tomographic reconstruction

    Elena Morotti, Davide Evangelista, and Elena Loli Piccolomini. A green prospective for learned post-processing in sparse-view tomographic reconstruction. Journal of Imaging, 7(8):139, 2021

    2021

  37. [45]

    Robust non-convex model-based approach for deep learning-based image processing

    Elena Morotti, Davide Evangelista, and Elena Loli Piccolomini. Robust non-convex model-based approach for deep learning-based image processing. In International Conference on Numerical Computations: Theory and Algorithms, pages 360–367. Springer, 2023

    2023

  38. [46]

    Minimizing nonconvex functions for sparse vector reconstruction

    Nasser Mourad and James P Reilly. Minimizing nonconvex functions for sparse vector reconstruction. IEEE Transactions on Signal Processing, 58(7):3485–3496, 2010

    2010

  39. [47]

    On iteratively reweighted algorithms for nonsmooth nonconvex optimization in computer vision

    Peter Ochs, Alexey Dosovitskiy, Thomas Brox, and Thomas Pock. On iteratively reweighted algorithms for nonsmooth nonconvex optimization in computer vision. SIAM Journal on Imaging Sciences , 8(1):331–372, 2015

    2015

  40. [48]

    Why do commercial ct scanners still employ traditional, filtered back-projection for image reconstruction? Inverse problems, 25(12):123009, 2009

    Xiaochuan Pan, Emil Y Sidky, and Michael Vannier. Why do commercial ct scanners still employ traditional, filtered back-projection for image reconstruction? Inverse problems, 25(12):123009, 2009

    2009

  41. [49]

    U-net: Convolutional networks for biomedical image seg- mentation

    Olaf Ronneberger, Philipp Fischer, and Thomas Brox. U-net: Convolutional networks for biomedical image seg- mentation. In Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, par...

    2015

  42. [50]

    Sidky and et al

    Emil Y . Sidky and et al. Constrained TpV minimization for enhanced exploitation of gradient sparsity: Appli- cation to CT image reconstruction. IEEE Journal of Translational Engineering in Health and Medicine, 2014

    2014

  43. [51]

    Sidky and et al

    Emil Y . Sidky and et al. Do CNNs solve the CT inverse problem?IEEE Transactions on Biomedical Engineering, 2021

    2021

  44. [52]

    Convex optimization problem prototyping for image reconstruction in computed tomography with the Chambolle–Pock algorithm

    Emil Y Sidky, Jakob H Jørgensen, and Xiaochuan Pan. Convex optimization problem prototyping for image reconstruction in computed tomography with the Chambolle–Pock algorithm. Physics in Medicine & Biology , 2012

    2012

  45. [53]

    Relaxed conditions for sparse signal recovery with general concave priors

    Joshua Trzasko and Armando Manduca. Relaxed conditions for sparse signal recovery with general concave priors. IEEE Transactions on Signal Processing, 57(11):4347–4354, 2009

    2009

  46. [54]

    Fast and flexible x-ray tomography using the astra toolbox

    Wim Van Aarle, Willem Jan Palenstijn, Jeroen Cant, Eline Janssens, Folkert Bleichrodt, Andrei Dabravolski, Jan De Beenhouwer, K Joost Batenburg, and Jan Sijbers. Fast and flexible x-ray tomography using the astra toolbox. Optics express, 24(22):25129–25147, 2016

    2016

  47. [55]

    The astra toolbox: A platform for advanced algorithm development in electron tomography

    Wim Van Aarle, Willem Jan Palenstijn, Jan De Beenhouwer, Thomas Altantzis, Sara Bals, K Joost Batenburg, and Jan Sijbers. The astra toolbox: A platform for advanced algorithm development in electron tomography. Ultramicroscopy, 157:35–47, 2015

    2015

  48. [56]

    Plug-and-play priors for model based reconstruction

    Singanallur V Venkatakrishnan, Charles A Bouman, and Brendt Wohlberg. Plug-and-play priors for model based reconstruction. In 2013 IEEE global conference on signal and information processing , pages 945–948. IEEE, 2013

    2013

  49. [57]

    A deep residual architecture for skin lesion segmentation

    GM Venkatesh, YG Naresh, Suzanne Little, and Noel E O’Connor. A deep residual architecture for skin lesion segmentation. In OR 2.0 Context-Aware Operating Theaters, Computer Assisted Robotic Endoscopy, Clinical Image-Based Procedures, and Skin Image Analysis: First Internation...

    2018

  50. [58]

    Admm-based deep reconstruction for limited-angle ct

    Jiaxi Wang, Li Zeng, Chengxiang Wang, and Yumeng Guo. Admm-based deep reconstruction for limited-angle ct. Physics in Medicine & Biology, 64(11):115011, 2019

    2019

  51. [59]

    Multiscale structural similarity for image quality assessment

    Zhou Wang, Eero P Simoncelli, and Alan C Bovik. Multiscale structural similarity for image quality assessment. In The Thrity-Seventh Asilomar Conference on Signals, Systems & Computers, 2003 , volume 2, pages 1398–

    2003

  52. [60]

    Iterative reweighted\ell 1 and\ell 2 methods for finding sparse solutions

    David Wipf and Srikantan Nagarajan. Iterative reweighted\ell 1 and\ell 2 methods for finding sparse solutions. IEEE Journal of Selected Topics in Signal Processing, 4(2):317–329, 2010

    2010

  53. [61]

    Deep convolutional framelets: A general deep learning framework for inverse problems

    Jong Chul Ye, Yoseob Han, and Eunju Cha. Deep convolutional framelets: A general deep learning framework for inverse problems. SIAM Journal on Imaging Sciences, 11(2):991–1048, 2018

    2018

  54. [62]

    Domain generalization for medical image analysis: A survey

    Jee Seok Yoon, Kwanseok Oh, Yooseung Shin, Maciej A Mazurowski, and Heung-Il Suk. Domain generalization for medical image analysis: A survey. arXiv preprint arXiv:2310.08598, 2023

    2023 arXiv

  55. [63]

    Low-dose ct via deep cnn with skip connection and network-in-network

    Chenyu You, Linfeng Yang, Yi Zhang, and Ge Wang. Low-dose ct via deep cnn with skip connection and network-in-network. In Developments in X-Ray tomography XII, volume 11113, pages 429–434. SPIE, 2019

    2019

  56. [64]

    Road extraction by deep residual u-net

    Zhengxin Zhang, Qingjie Liu, and Yunhong Wang. Road extraction by deep residual u-net. IEEE Geoscience and Remote Sensing Letters, 15(5):749–753, 2018

    2018

  57. [65]

    Hybrid skip: A biologically inspired skip connection for the unet architecture

    Nikolaos Zioulis, Georgios Albanis, Petros Drakoulis, Federico Alvarez, Dimitrios Zarpalas, and Petros Daras. Hybrid skip: A biologically inspired skip connection for the unet architecture. IEEE Access, 10:53928–53939, 2022

    2022

  58. [66]

    Fondo per il Programma Nazionale di Ricerca e Progetti di Rilevante Inter- esse Nazionale (PRIN)

    Hui Zou and Runze Li. One-step sparse estimates in nonconcave penalized likelihood models. Annals of statis- tics, 36(4):1509, 2008. Acknowledgement The author is partially supported by the “Fondo per il Programma Nazionale di Ricerca e Progetti di Rilevante Inter- esse Nazion...

    2008

  59. [1402]

    21 A PREPRINT - MAY 14, 2025

    Ieee, 2003. 21 A PREPRINT - MAY 14, 2025

    2003

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.