Pith. sign in

REVIEW 5 major objections 4 minor 46 references

Direct Dual-Energy CT Material Decomposition using Model-based Denoising Diffusion Model

T0 review · 5 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A learned inverse material model turns dual-energy CT projections directly into water and bone density maps.

desk verdict Plausible integration of a learned sinogram-domain decomposition with a diffusion prior, but the consistency claim is overreaching and Eq. (31) has a real dimensional error. read the letter →

arxiv 2507.18012 v1 pith:CLJK423Q submitted 2025-07-24 eess.IV cs.CVphysics.med-ph

Hang Xu , Alexandre Bousse , Alessandro Perelli This is my paper · ORCID
classification eess.IVcs.CVphysics.med-ph
keywords dual-energyCTmaterialdecompositiondenoisingdiffusionmodelprojection-domainsparse-viewmodel-basediterativereconstructionscore-basedpriorquantitative
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 sets out to show that dual-energy CT material decomposition can skip image-domain post-processing entirely: a learned network maps the low- and high-energy sinograms directly into water and bone density images, with a denoising diffusion model supplying the image prior and a data-consistency term keeping the result tied to the measured projections. The proposed DEcomp-MoD pipeline is trained and evaluated on synthetic low-dose, sparse-view DECT sinograms generated from real patient CT images, and the reported numbers put it ahead of both a supervised one-step network and a score-based diffusion baseline. If the claim holds, the practical payoff is a fast, calibration-free route from raw DECT projections to quantitative material maps that handles beam hardening and sparse-view artifacts without explicit inversion of the polyenergetic model.

What carries the argument

The load-bearing object is the learned decomposition map $P_\theta \approx h^{-1}$ embedded directly inside the negative log-likelihood term (Eq. 28). This substitution is what linearizes the otherwise nonlinear DECT model: instead of inverting the polyenergetic Beer-Lambert relation numerically or by phantom calibration, the U-Net supplies material sinograms $P_\theta(y)$, and the data-consistency update becomes a weighted quadratic minimization solved with conjugate gradient. Around it, the diffusion prior is applied through a proximal denoising step matching the DDIM reverse update, with parameters $\lambda$ and $\xi$ trading prior strength against measurement fidelity. The HQS split makes the two learning modules independent: the U-Net works in sinogram space, the diffusion model in image space, and the CG solve couples them.

What would settle it

Take a calibrated dual-energy CT system with known water and bone phantom inserts, acquire real low-dose and sparse-view sinograms, run the pre-trained DEcomp-MoD without any fine-tuning, and compare the recovered material densities against the known insert values; a systematic bias that grows with tube voltage, filtration, or added phantom material would show the learned $h^{-1}$ was tied to the simulation rather than to the physical scanner.

Watch

Extended reading notes

Core claim

DEcomp-MoD's central claim is that the hard nonlinear inverse of the polyenergetic DECT measurement model, $h^{-1}$, can be replaced by a supervised U-Net $P_\theta$ trained on simulated (energy sinogram, material sinogram) pairs, turning the data-consistency sub-problem into a quadratic least-squares fit. Around that fit, an unconditional denoising diffusion model acts as a plug-and-play prior on material images, and the two are coupled through half-quadratic splitting with conjugate-gradient updates. The algorithm iterates a DDIM-style reverse diffusion step followed by a projection-domain data-consistency correction, which the paper argues guarantees consistency of the results in the sense that every update pulls the material image back toward the measured sinograms. Quantitative results on simulated low-dose patient data give water PSNR 29.14 dB and bone PSNR 33.88 dB, above the compared baselines.

Load-bearing premise

The whole pipeline depends on the simulated U-Net $P_\theta$ being a faithful substitute for the true inverse material decomposition on real scanner data; the experiments only test $P_\theta$ on synthetic sinograms generated with the same spectra, segmentation thresholds, and attenuation tables used in training, so a real scanner with different beam quality, scatter, or detector response could break the consistency guarantee.

Editorial extensions

If this is right

  • Material decomposition no longer needs a separate reconstruction step: the energy sinograms are mapped to material sinograms and then to images through a closed-form quadratic solve, so beam-hardening and noise correlations from the polyenergetic model are handled in one pass.
  • The learned $P_\theta$ removes the need for explicit calibration of $h^{-1}$, since the U-Net is trained from simulated spectral data instead of a measured calibration phantom.
  • On the simulated low-dose, sparse-view test set, the method reports water PSNR 29.14 dB and bone PSNR 33.88 dB, outperforming both the supervised one-step network and the score-based diffusion baseline.
  • The diffusion prior generalizes across sampling angles: trained at 180 angles, DEcomp-MoD keeps higher PSNR and SSIM than the baselines from 60 to 360 angles without retraining.
  • With 100 sampling steps a 256x256 material image is produced in about 9.4 seconds, which is competitive with unrolled iterative methods and far faster than the manifold-constraint diffusion baseline.

Reading between the lines

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

  • Not stated in the paper: because the diffusion prior is trained on material images rather than sinograms, the same prior could be reattached to different scanner geometries or spectra by retraining only the smaller U-Net; the authors note a future unsupervised direction, but the modular split makes this concrete.
  • Not stated in the paper: the covariance matrix $B_n$ computed by backpropagating through $P_\theta$ could be interpreted as an uncertainty map for the decomposition, offering a per-pixel confidence estimate that the current evaluation does not report.
  • Not stated in the paper: replacing the supervised U-Net with a self-supervised estimator trained on paired noisy sub-sampled sinograms would remove the dependence on simulated ground-truth material sinograms and directly test whether the consistency guarantee transfers to real scanner data.
  • Not stated in the paper: if $P_\theta$ were swapped for a differentiable calibration-based decomposition, the same diffusion/HQS loop would still run, making the framework a testbed for whether learned decompositions beat classical polynomial fits independently of the diffusion prior.
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

5 major / 4 minor

Summary. DEcomp-MoD is a two-module deep-learning pipeline for dual-energy CT material decomposition. A U-Net Pθ is trained in a supervised way on simulated dual-energy sinograms to approximate the inverse material decomposition h^{-1}, mapping the two energy sinograms into material sinograms. The second module is a pre-trained DDPM/DDIM score-based prior on material images. At inference, HQS alternates between a diffusion denoising step and a quadratic data-consistency step in which the material sinogram produced by Pθ(y) is fitted to Ax under a per-ray covariance B_n, solved with CG. Experiments on synthetic low-dose, sparse-view DECT sinograms simulated from the AAPM low-dose CT dataset compare water/bone PSNR and SSIM against FBP, MCG diffusion, and DIRECT-Net, and include an angular-undersampling generalization study and a parameter study of λ and ξ.

Significance. If fully supported, the method would have practical value: decoupled training avoids end-to-end unrolling, the diffusion prior can be reused across acquisition settings, and the projection-domain decomposition makes the data-consistency subproblem quadratic and relatively cheap (about 10 s per image). The paper gives a reasonably complete description of the simulation pipeline, the network architectures, and Algorithm 1, which helps reproducibility. However, the central 'guarantees consistency' claim is not supported by Eq. (28), the comparison claims are weakened by test-set tuning and by the single-slice MCG baseline, and the closed-form update in Eq. (31) is not reproducible as written.

major comments (5)
  1. [V-A, Eq. (28)] The data-consistency term fθ(y,x) in Eq. (28) is ||Pθ(y) − Ax||^2_B, not a comparison of x with the measured sinogram y; it only enforces consistency with the output of the learned U-Net. Since Pθ is trained and tested on Spekpy-simulated pairs from a single simulation pipeline (Section VI-A2) and is never validated against a calibrated h^{-1} or any real DECT data, the Abstract and Section I-A claim that DEcomp-MoD 'guarantees consistency of the results' is an overclaim. I recommend either augmenting the inference objective with the spectral forward-model term y ≈ h(Ax), or rephrasing the claim as consistency with the learned projection-domain decomposition and adding a domain-shift experiment (e.g., testing on spectra or filtration different from training).
  2. [VI-F1] The default values λ = 0.001 and ξ = 1.0 are selected 'based on the results in section VI-F1', which are obtained by examining qualitative and quantitative testing performance on the same test set used for Table I and Fig. 8. This is effectively tuning on the test set and makes the reported PSNR/SSIM and the comparisons with FBP, MCG, and DIRECT-Net optimistic. Please introduce a separate validation set, or report both validation and test numbers, before computing the final comparisons.
  3. [V-A, Eq. (31)] Equation (31) is dimensionally inconsistent: B ∈ R^{N×2} is defined as the concatenation of the diagonal entries of B_n, while A ∈ R^{N×M} with M = 256^2 in the experiments, so the elementwise product B ⊙ A is not defined. The closed-form solution of the weighted least-squares problem (30) requires a block-diagonal weight matrix acting on vec(Ax) (e.g., a matrix with blocks B_n along the ray index), leading to normal equations of the form (Ãᵀ D Ã + μI)x = Ãᵀ D p + μz. Please give the correct matrix form and state what was actually implemented; as written, Algorithm 1 is not reproducible.
  4. [V-A, Eq. (29)] The covariance formula in Eq. (29) appears to have the Jacobian orientation reversed. With W_n the inverse covariance of y_n and Pθ(y)_n approximating h^{-1}(y)_n, the first-order covariance of the decomposed sinogram should scale as ∇Pθ(y)_n W_n^{-1} ∇Pθ(y)_nᵀ (or the equivalent forward-model form), whereas [∇Pθ(y)_n]^{-1} W_n [∇Pθ(y)_n]^{-T} would make the weight increase with photon count, which is unphysical. Please correct the formula or provide the derivation from [17] showing the intended convention.
  5. [VI-D, Table I] The MCG diffusion baseline is evaluated on a single slice (as acknowledged in the text below Table I), and its PSNR/SSIM entries have no standard deviation. The claim that DEcomp-MoD outperforms 'state-of-the-art unsupervised score-based model' is therefore not established over the test population. Please run MCG on the full test set or, if the cost is prohibitive, clearly state that the comparison is illustrative on one slice and remove the global comparative claim.
minor comments (4)
  1. [VI-B1 / VII-3] The description of Pθ is inconsistent: Eq. (26) and Algorithm 1 treat Pθ as mapping R^{N×2} to R^{N×2}, but Section VI-B1 says the U-Net output is a one-channel material sinogram, and Section VII-3 says only one material can be generated at a time. Please clarify whether two networks are trained or whether a two-channel output is used.
  2. [VI-E] In Section VI-E, 'the CT data is acquired at difference conditions' contains a typo ('difference' should be 'different'), and the sentence 'with fixed photon counts 2·10^6 and different number of sampling angles' is missing a verb; the paragraph should be edited for clarity.
  3. [V-A after Eq. (31)] There is a notation inconsistency after Eq. (31), where G is defined as Aᵀ(B ⊙ A) + λI but the preceding equation uses μ; the text should use one symbol consistently (μ or λ).
  4. [VI-C] The default CG iterations are fixed to 10, but the CG tolerance or stopping criterion is not stated; since the update (35b) is solved approximately at every diffusion step, please report the stopping rule or tolerance used.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the learned surrogate P_theta for h^{-1} is explicitly disclosed and trained by a supervised loss, so the inference consistency term is an internal modeling choice, not a hidden equivalence.

full rationale

The only step that could look circular is the substitution in Eq. (26), '[[pn = h−1(y)n ≈ P θ(y)n', followed by Eq. (28), where the inference data-consistency term is ‖Pθ(y)n − [Ax]n‖2_Bn. This is not a hidden identity: Pθ is an explicit trainable U-Net, fitted by Eq. (27) (min ½‖Pθ(y)−p*‖2 with p*=Ax*) on simulated training pairs. The inference objective therefore guarantees consistency with the learned pseudo-sinogram Pθ(y), not directly with y through Beer's law as in Eq. (10); that mismatch is an external-validity limitation of the claim that the method 'guarantees consistency of the results', but it is disclosed in the method itself, not smuggled in as an equality by construction. The final material image is not simply Pθ(y): the HQS splitting in Eq. (24) includes a diffusion prior and a CG solve, so the output is the solution of a nontrivial optimization, not the training label itself. The author-inclusive citations [26] and [31] are related-work references on diffusion-based spectral CT and are not load-bearing for the core derivation; no uniqueness theorem is imported, and the comparisons against DIRECT-Net and MCG are external baselines. Thus no step of the derivation reduces to its own input by definition, and the paper should not be scored as circular despite the legitimate concern that the learned surrogate may not transfer to real scanner data.

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

The central reconstruction depends on two fitted networks, the decomposition U-Net and the diffusion prior, both trained on the same simulation pipeline, plus hyperparameters tuned on testing data. No new physical entities are introduced.

free parameters (5)
  • lambda = 0.001
    Regularization weight balancing data consistency and diffusion prior; chosen empirically based on testing performance in Section VI-F1.
  • xi = 1.0
    Noise trade-off parameter in the DDIM reverse sampling step; chosen empirically based on testing performance in Section VI-F1.
  • CG iterations = 10
    Fixed for all experiments; no ablation study is provided to justify this value.
  • Sampling iterations T = 100
    Default number of reverse diffusion steps; Table II shows PSNR varies significantly with this setting, so the chosen value affects results.
  • Material segmentation thresholds = not specified
    Bone and water labels are generated by thresholding [40], but the actual threshold values are not reported; these thresholds determine the training labels and the ground truth used for evaluation.
assumptions (5)
  • domain assumption The two-basis-material decomposition with known mass attenuation functions for water and bone is sufficient to represent the energy-dependent attenuation (Eq. 9).
    This is the standard DECT assumption but limits applicability to materials that are not a linear combination of the chosen bases.
  • domain assumption A U-Net P_theta trained on simulated pairs (y, p*) approximates the true inverse h^{-1} on unseen data (Eq. 26, Section V-A).
    The data consistency term at inference uses P_theta(y) as if it were the exact material sinogram; no real-data calibration is performed.
  • domain assumption The diffusion model trained on threshold-segmented material images from the AAPM dataset is a valid prior for generalized clinical material images.
    Only one dataset and one segmentation method are used; generalization to other scanners, body regions, or dose levels is unproven.
  • ad hoc to paper First-order Taylor expansion through P_theta gives a valid covariance B_n for the transformed measurements (Eq. 29).
    The linearization is asserted and its accuracy is not evaluated in the paper.
  • standard math The Poisson-distributed photon counts can be approximated by an anisotropic Gaussian with diagonal covariance (Section III).
    This is a widely used low-count approximation in CT reconstruction, but it is an approximation, not an exact model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Direct Dual-Energy CT Material Decomposition using Model-based Denoising Diffusion Model." pith.science (2026). https://pith.science/paper/CLJK423Q

@misc{pith2026250718012,
  author       = {Pith},
  title        = {Pith review of: Direct Dual-Energy CT Material Decomposition using Model-based Denoising Diffusion Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CLJK423Q}},
  note         = {Machine review of arXiv:2507.18012}
}
read the original abstract

Dual-energy X-ray Computed Tomography (DECT) constitutes an advanced technology which enables automatic decomposition of materials in clinical images without manual segmentation using the dependency of the X-ray linear attenuation with energy. However, most methods perform material decomposition in the image domain as a post-processing step after reconstruction but this procedure does not account for the beam-hardening effect and it results in sub-optimal results. In this work, we propose a deep learning procedure called Dual-Energy Decomposition Model-based Diffusion (DEcomp-MoD) for quantitative material decomposition which directly converts the DECT projection data into material images. The algorithm is based on incorporating the knowledge of the spectral DECT model into the deep learning training loss and combining a score-based denoising diffusion learned prior in the material image domain. Importantly the inference optimization loss takes as inputs directly the sinogram and converts to material images through a model-based conditional diffusion model which guarantees consistency of the results. We evaluate the performance with both quantitative and qualitative estimation of the proposed DEcomp-MoD method on synthetic DECT sinograms from the low-dose AAPM dataset. Finally, we show that DEcomp-MoD outperform state-of-the-art unsupervised score-based model and supervised deep learning networks, with the potential to be deployed for clinical diagnosis.

Figures

Figures reproduced from arXiv: 2507.18012 by the authors.

Figure 1
Figure 1. Schematic of the diffusion model. The forward pro [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Schematic of the U-Net which implements the [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Diagram for the supervised training of the projection [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: DEcomp-MoD algorithmic workflow for material decomposition. The energy sinogram [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: The schematic of DECT energy sinograms sim [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Qualitative decomposition results for soft tissue and bone using: (a) FBP, (b) MCG diffusion, (c) DIRECT-Net, (d) [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Qualitative results of FBP, DIRECT-Net and DEcomp-MoD using measurement projections acquired from different [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 8
Figure 8. Figure 8: Generalization of FBP, DIRECT-Net and DEcomp-MoD. The three algorithms are trained with data acquired by [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]
Figure 9
Figure 9. Figure 9: Qualitative results of DEcomp-MoD for different set of parameters [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Reverse inference process of DEcomp-MoD using different number of sampling iterations [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 44 canonical work pages

  1. [31]

    Spectral CT Two-step and One-step Material Decomposition using Diffusion Posterior Sampling

    C. Vazia, A. Bousse, J. Froment, B. Vedel, F. Vermet, Z. Wang, T. Dassow, J.-P. Tasu, and D. Visvikis, “Spectral CT two-step and one- step material decomposition using diffusion posterior sampling,” arXiv preprint arXiv:2403.10183, 2024

    work page Pith review arXiv 2024

  2. [32]

    Multi-Material Decomposition Using Spectral Diffusion Posterior Sampling

    X. Jiang, G. J. Gang, and J. W. Stayman, “Multi-material decom- position using spectral diffusion posterior sampling,” arXiv preprint arXiv:2408.01519, 2024

    work page Pith review arXiv 2024

  3. [17]

    Model-based iterative reconstruction for dual-energy X-ray CT using a joint quadratic likelihood model,

    R. Zhang, J.-B. Thibault, C. A. Bouman, K. D. Sauer, and J. Hsieh, “Model-based iterative reconstruction for dual-energy X-ray CT using a joint quadratic likelihood model,” IEEE transactions on medical imaging, vol. 33, no. 1, pp. 117–134, 2013

    work page 2013

  4. [1]

    A flexible method for multi-material decomposition of dual-energy CT images,

    P. R. Mendonc ¸a, P. Lamb, and D. V . Sahani, “A flexible method for multi-material decomposition of dual-energy CT images,” IEEE transactions on medical imaging , vol. 33, no. 1, pp. 99–116, 2013

    work page 2013

  5. [2]

    Quantitative imaging of element composition and mass fraction using dual-energy CT: Three-material decomposition,

    X. Liu, L. Yu, A. N. Primak, and C. H. McCollough, “Quantitative imaging of element composition and mass fraction using dual-energy CT: Three-material decomposition,” Medical physics, vol. 36, no. 5, pp. 1602–1609, 2009

    work page 2009

  6. [3]

    Quantitative dual-energy micro-CT with a photon-counting detector for material science and non-destructive testing,

    T. Sellerer, S. Ehn, K. Mechlem, M. Duda, M. Epple, P. B. No ¨el, and F. Pfeiffer, “Quantitative dual-energy micro-CT with a photon-counting detector for material science and non-destructive testing,” Plos one , vol. 14, no. 7, p. e0219659, 2019

    work page 2019

  7. [4]

    Dual energy computed tomography for explosive detection,

    Z. Ying, R. Naidu, and C. R. Crawford, “Dual energy computed tomography for explosive detection,” Journal of X-ray Science and Technology, vol. 14, no. 4, pp. 235–256, 2006. 13

    work page 2006

  8. [5]

    State of the art: dual-energy CT of the abdomen,

    D. Marin, D. T. Boll, A. Mileto, and R. C. Nelson, “State of the art: dual-energy CT of the abdomen,” Radiology, vol. 271, no. 2, pp. 327– 342, 2014

    work page 2014

Show all 46 references
  1. [6]

    First performance evaluation of a dual-source CT (DSCT) system,

    T. G. Flohr, C. H. McCollough, H. Bruder, M. Petersilka, K. Gruber, C. S¨uβ, M. Grasruck, K. Stierstorfer, B. Krauss, R. Raupach et al., “First performance evaluation of a dual-source CT (DSCT) system,” European radiology, vol. 16, pp. 256–268, 2006

    2006

  2. [7]

    Dual layer CT,

    A. Vlassenbroek, “Dual layer CT,” Dual energy CT in clinical practice , pp. 21–34, 2011

    2011

  3. [8]

    Energy-selective reconstructions in X- ray computerised tomography,

    R. E. Alvarez and A. Macovski, “Energy-selective reconstructions in X- ray computerised tomography,” Physics in Medicine & Biology , vol. 21, no. 5, p. 733, 1976

    1976

  4. [9]

    Material differentiation by dual energy CT: initial experience,

    T. R. Johnson, B. Krauss, M. Sedlmair, M. Grasruck, H. Bruder, D. Morhard, C. Fink, S. Weckbach, M. Lenhard, B. Schmidt et al. , “Material differentiation by dual energy CT: initial experience,” Euro- pean radiology , vol. 17, pp. 1510–1517, 2007

    2007

  5. [10]

    Image-based dual energy CT using optimized precorrection functions: A practical new approach of material decomposition in image domain,

    C. Maaß, M. Baer, and M. Kachelrieß, “Image-based dual energy CT using optimized precorrection functions: A practical new approach of material decomposition in image domain,” Medical physics , vol. 36, no. 8, pp. 3818–3829, 2009

    2009

  6. [11]

    Iterative image-domain decomposition for dual-energy CT,

    T. Niu, X. Dong, M. Petrongolo, and L. Zhu, “Iterative image-domain decomposition for dual-energy CT,” Medical physics , vol. 41, no. 4, p. 041901, 2014

    2014

  7. [12]

    Image domain dual material decomposition for dual-energy CT using butterfly network,

    W. Zhang, H. Zhang, L. Wang, X. Wang, X. Hu, A. Cai, L. Li, T. Niu, and B. Yan, “Image domain dual material decomposition for dual-energy CT using butterfly network,” Medical physics, vol. 46, no. 5, pp. 2037– 2051, 2019

    2019

  8. [13]

    Material decomposition from photon-counting CT using a convolutional neural network and energy-integrating CT training labels,

    R. Nadkarni, A. Allphin, D. P. Clark, and C. T. Badea, “Material decomposition from photon-counting CT using a convolutional neural network and energy-integrating CT training labels,” Physics in Medicine & Biology , vol. 67, no. 15, p. 155003, 2022

    2022

  9. [14]

    A material decom- position method for dual-energy CT via dual interactive Wasserstein generative adversarial networks,

    Z. Shi, H. Li, Q. Cao, Z. Wang, and M. Cheng, “A material decom- position method for dual-energy CT via dual interactive Wasserstein generative adversarial networks,” Medical Physics , vol. 48, no. 6, pp. 2891–2905, 2021

    2021

  10. [15]

    Systematic review on learning-based spectral CT,

    A. Bousse, V . S. S. Kandarpa, S. Rit, A. Perelli, M. Li, G. Wang, J. Zhou, and G. Wang, “Systematic review on learning-based spectral CT,” IEEE transactions on radiation and plasma medical sciences , vol. 8, no. 2, pp. 113–137, 2023

    2023

  11. [16]

    An accurate method for direct dual- energy calibration and decomposition,

    H. N. Cardinal and A. Fenster, “An accurate method for direct dual- energy calibration and decomposition,” Medical physics , vol. 17, no. 3, pp. 327–341, 1990

    1990

  12. [18]

    Spectral angiography material decomposition using an empirical forward model and a dictionary-based regularization,

    K. Mechlem, T. Sellerer, S. Ehn, D. M ¨unzel, E. Braig, J. Herzen, P. B. No¨el, and F. Pfeiffer, “Spectral angiography material decomposition using an empirical forward model and a dictionary-based regularization,” IEEE transactions on medical imaging , vol. 37, pp. 2298–2309, 2018

    2018

  13. [19]

    Multi-material decomposition using statistical image reconstruction for spectral CT,

    Y . Long and J. Fessler, “Multi-material decomposition using statistical image reconstruction for spectral CT,” IEEE transactions on medical imaging, vol. 33, no. 8, pp. 1614–1626, 2014

    2014

  14. [20]

    Joint statistical iterative material image reconstruction for spectral computed tomography using a semi-empirical forward model,

    K. Mechlem, S. Ehn, T. Sellerer, E. Braig, D. M ¨unzel, F. Pfeiffer, and P. No¨el, “Joint statistical iterative material image reconstruction for spectral computed tomography using a semi-empirical forward model,” vol. 37, pp. 68–80, 2018

    2018

  15. [21]

    An algorithm for constrained one-step inversion of spectral CT data,

    R. Foygel Barber, E. Sidky, T. Gilat Schmidt, and X. Pan, “An algorithm for constrained one-step inversion of spectral CT data,” Phys Med Biol , vol. 61, no. 10, pp. 3784–3818, May 2016

    2016

  16. [22]

    Physics-/model-based and data-driven methods for low-dose computed tomography: A survey,

    W. Xia, H. Shan, G. Wang, and Y . Zhang, “Physics-/model-based and data-driven methods for low-dose computed tomography: A survey,” IEEE Signal Processing Magazine , vol. 40, no. 2, pp. 89–100, 2023

    2023

  17. [23]

    DIRECT-Net: A unified mutual- domain material decomposition network for quantitative dual-energy CT imaging,

    T. Su, X. Sun, J. Yang, D. Mi, Y . Zhang, H. Wu, S. Fang, Y . Chen, H. Zheng, D. Liang, and Y . Ge, “DIRECT-Net: A unified mutual- domain material decomposition network for quantitative dual-energy CT imaging,” Med Phys , vol. 49, no. 2, pp. 917–934, 2022

    2022

  18. [24]

    Regularization by denoising sub-sampled newton method for spectral CT multi-material decomposition,

    A. Perelli and M. Andersen, “Regularization by denoising sub-sampled newton method for spectral CT multi-material decomposition,” Philo- sophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , vol. 379, no. 2200, p. 20200191, 2021

    2021

  19. [25]

    Iterative material decomposition for spectral CT using self-supervised Noise2Noise prior,

    W. Fang, D. Wu, K. Kim, M. Kalra, R. Singh, L. Li, and Q. Li, “Iterative material decomposition for spectral CT using self-supervised Noise2Noise prior,” Phys Med Biol , vol. 66, no. 15, p. 155013, 2021

    2021

  20. [26]

    Material decomposition in photon-counting computed tomography with diffusion models: Com- parative study and hybridization with variational regularizers,

    C. Vazia, T. Dassow, A. Bousse, J. Froment, B. Vedel, F. Vermet, A. Perelli, J.-P. Tasu, and D. Visvikis, “Material decomposition in photon-counting computed tomography with diffusion models: Com- parative study and hybridization with variational regularizers,” arXiv preprint ...

    2025

  21. [27]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in neural information processing systems , vol. 33, 2020

    2020

  22. [28]

    Palette: Image-to-image diffusion models,

    C. Saharia, W. Chan, H. Chang, C. Lee, J. Ho, T. Salimans, D. Fleet, and M. Norouzi, “Palette: Image-to-image diffusion models,” in ACM SIGGRAPH 2022 Conference Proceedings , 2022, pp. 1–10

    2022

  23. [29]

    Denoising diffusion restoration models,

    B. Kawar, M. Elad, S. Ermon, and J. Song, “Denoising diffusion restoration models,” Advances in Neural Information Processing Sys- tems, vol. 35, pp. 23 593–23 606, 2022

    2022

  24. [30]

    Solving inverse problems in medical imaging with score-based generative models,

    Y . Song, L. Shen, L. Xing, and S. Ermon, “Solving inverse problems in medical imaging with score-based generative models,” in NeurIPS 2021 Workshop on Deep Learning and Inverse Problems , 2021

    2021

  25. [33]

    Diffusion models as plug-and-play priors,

    A. Graikos, N. Malkin, N. Jojic, and D. Samaras, “Diffusion models as plug-and-play priors,” Advances in Neural Information Processing Systems, vol. 35, pp. 14 715–14 728, 2022

    2022

  26. [34]

    Denoising diffusion models for plug-and-play image restoration,

    Y . Zhu, K. Zhang, J. Liang, J. Cao, B. Wen, R. Timofte, and L. Van Gool, “Denoising diffusion models for plug-and-play image restoration,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 1219–1229

    2023

  27. [35]

    Low-dose CT for the detection and classification of metastatic liver lesions: results of the 2016 low dose CT grand challenge,

    C. H. McCollough, A. C. Bartley, R. E. Carter, B. Chen, T. A. Drees, P. Edwards, D. R. Holmes III, A. E. Huang, F. Khan, S. Leng et al. , “Low-dose CT for the detection and classification of metastatic liver lesions: results of the 2016 low dose CT grand challenge,” Medical ph...

    2016

  28. [36]

    A recursive filter for noise reduction in statistical iterative tomographic imaging,

    J.-B. Thibault, C. A. Bouman, K. D. Sauer, and J. Hsieh, “A recursive filter for noise reduction in statistical iterative tomographic imaging,” in Computational Imaging IV , vol. 6065. SPIE, 2006, pp. 264–273

    2006

  29. [37]

    Denoising diffusion implicit models,

    J. Song, C. Meng, and S. Ermon, “Denoising diffusion implicit models,” in International Conference on Learning Representations , 2020

    2020

  30. [38]

    Generative modeling by estimating gradients of the data distribution,

    Y . Song and S. Ermon, “Generative modeling by estimating gradients of the data distribution,” in Advances in Neural Information Processing Systems, 2019, pp. 11 895–11 907

    2019

  31. [39]

    Conjugate gradient method,

    J. L. Nazareth, “Conjugate gradient method,” Wiley Interdisciplinary Reviews: Computational Statistics , vol. 1, no. 3, pp. 348–353, 2009

    2009

  32. [40]

    Improving PET/CT attenuation correction with iterative CT beam hardening correction,

    M. Kachelrieb and W. A. Kalender, “Improving PET/CT attenuation correction with iterative CT beam hardening correction,” in IEEE Nuclear Science Symposium Conference Record , vol. 4. IEEE, 2005

    2005

  33. [41]

    Tomosipo: Fast, flexible, and convenient 3D tomography for complex scanning geometries in Python,

    A. Hendriksen, D. Schut, W. J. Palenstijn, N. Vigan `o, J. Kim, D. Pelt, T. van Leeuwen, and K. J. Batenburg, “Tomosipo: Fast, flexible, and convenient 3D tomography for complex scanning geometries in Python,” Optics Express, 2021

    2021

  34. [42]

    Spekpy v2. 0—a software toolkit for modeling X-ray tube spectra,

    G. Poludniowski, A. Omar, R. Bujila, and P. Andreo, “Spekpy v2. 0—a software toolkit for modeling X-ray tube spectra,” Medical Physics , vol. 48, no. 7, pp. 3630–3637, 2021

    2021

  35. [43]

    Tables of X-ray mass attenuation coefficients and mass energy-absorption coefficients 1 kev to 20 mev for elements Z = 1 to 92 and 48 additional substances of dosimetric interest,

    J. H. Hubbell and S. M. Seltzer, “Tables of X-ray mass attenuation coefficients and mass energy-absorption coefficients 1 kev to 20 mev for elements Z = 1 to 92 and 48 additional substances of dosimetric interest,” National Inst. of Standards and Technology-PL, Gaithersburg, T...

    1995

  36. [44]

    Equivariant imaging: Learning beyond the range space,

    D. Chen, J. Tachella, and M. E. Davies, “Equivariant imaging: Learning beyond the range space,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2021, pp. 4379–4388

    2021

  37. [45]

    Pytorch: An imperative style, high-performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga et al. , “Pytorch: An imperative style, high-performance deep learning library,” Advances in neural information processing systems , vol. 32, 2019

    2019

  38. [46]

    Improving diffusion models for inverse problems using manifold constraints,

    H. Chung, B. Sim, and J. C. Ye, “Improving diffusion models for inverse problems using manifold constraints,” in Advances in Neural Information Processing Systems , 2022

    2022

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.