Pith. sign in

REVIEW 5 major objections 8 minor 73 references

FgC2F-UDiff: Frequency-guided and Coarse-to-fine Unified Diffusion Model for Multi-modality Missing MRI Synthesis

T0 review · 5 major / 8 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read FgC2F-UDiff proposes a single conditional diffusion model that, guided by low- and high-frequency priors in a coarse-to-fine denoising schedule, synthesizes any missing MRI modality from any combination of available ones.

desk verdict A plausible unified missing-MRI diffusion model whose key frequency-staging premise and ablation attribution are under-supported; worth revising, not rejecting. read the letter →

arxiv 2501.03526 v1 pith:SASHRVXH submitted 2025-01-07 eess.IV cs.CVcs.LG

Xiaojiao Xiao , Qinmin Vivian Hu , Guanghui Wang This is my paper · ORCID
classification eess.IVcs.CVcs.LG
keywords MRIsynthesisdiffusionmodelmissingmodalityfrequencyguidancecoarse-to-finedenoisingmany-to-manyBraTS2021IXIdataset
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

This paper tries to establish that one unified diffusion model can synthesize any missing MRI modality from whatever combination of available modalities remains, a many-to-many problem that current methods handle only with per-task networks or unstable GANs. The proposed model, FgC2F-UDiff, splits the reverse denoising process into a coarse stage and a fine stage, guiding the first with low-frequency anatomical structure and the second with high-frequency texture and edges. On BraTS 2021 and IXI brain MRI datasets, the paper reports higher PSNR, SSIM, and LPIPS and lower FID than seven task-specific and unified baselines across one-to-one, many-to-one, and many-to-many tasks. It also reports that the same gains hold with only 200 denoising steps, making diffusion-based synthesis fast enough for practical use. If the claims hold, a single trained model could replace many specialized synthesis networks in clinical and research settings.

What carries the argument

The load-bearing mechanism is the coupling of a coarse-to-fine reverse denoising schedule with frequency-guided conditioning. The Coarse-to-fine Unified Network (CUN) splits denoising into stages from $T$ to $T/2-1$ and from $T/2$ to $0$, mirroring the claimed global-to-detail property of diffusion. The Frequency-guided Collaborative Strategy (FCS) supplies a low-frequency prior image in the coarse stage and a high-frequency prior in the fine stage, chosen by a fixed left-to-right or right-to-left scan over the available modalities, with Gaussian low-pass and high-pass filters generating the priors. The Specific-acceleration Hybrid Mechanism (SHM) adds curriculum learning across missing-modality difficulty, the coarse-to-fine split itself, and dynamic frequency selection, allowing $T=200$ timesteps with roughly two seconds per image. All of this feeds a five-channel UNet that takes noisy target channels, available source channels, and one frequency channel, and outputs all target modalities in a single pass.

What would settle it

Run the same architecture with the frequency schedule swapped, so high-frequency guidance is used in the coarse stage and low-frequency guidance in the fine stage, and move the split from $T/2$ to $T/4$ and $3T/4$ on BraTS 2021. If PSNR and SSIM do not drop for the swapped order, or barely change with the split point, the coarse-to-fine frequency guidance is not the cause of the reported improvement.

Watch

Extended reading notes

Core claim

The paper's central discovery is that conditioning the two halves of diffusion denoising on separate frequency bands produces higher-fidelity cross-modality synthesis. FgC2F-UDiff divides the reverse Markov chain at $T/2$: the coarse stage takes a low-pass-filtered image from the first available modality in the order T1, T2, FLAIR, T1ce, and the fine stage takes a high-pass-filtered image from the first available modality in the reverse order. Both the available source modalities and the chosen frequency image enter a UNet as a five-channel condition, and the network outputs all missing target modalities in a single pass. Curriculum learning, ordered from one missing modality to three, and the coarse-to-fine split itself form the acceleration mechanism that allows $T=200$ with roughly two seconds per image. The reported experiments, with statistical significance at $p<0.05$, show the best metrics against pix2pix, pGAN, LDM, CoLa-Diff, MM-GAN, ResViT, and Uni-GAN on two datasets, and ablations attribute the gain to the coarse-to-fine network, the frequency guidance, and the hybrid acceleration mechanism.

Load-bearing premise

The load-bearing premise is that diffusion denoising really does progress from global structure to fine detail, so that splitting the reverse process at exactly $T/2$ and guiding the first half with low-frequency and the second half with high-frequency information is the correct and beneficial design; the paper supports this only with a qualitative inspection of its own denoised images and never varies the split or swaps the frequency order.

Editorial extensions

If this is right

  • A single trained network can handle many-to-many synthesis, covering all tested availability patterns of the four BraTS modalities without per-task retraining.
  • More available input modalities directly raise synthesis quality, with reported BraTS PSNR rising from 26.13 dB with only T1 to 29.43 dB when three modalities are available.
  • Diffusion sampling can be accelerated to 200 steps, about 1.5 to 2 seconds per image, without losing quality relative to latent diffusion baselines that need 800 to 1000 steps.
  • Frequency-guided coarse-to-fine conditioning also improves one-to-one tasks, with the paper reporting gains of roughly 0.85 dB over CoLa-Diff for T1 to T1ce synthesis on BraTS.
  • Curriculum learning from one missing modality to three missing modalities helps the model adapt to arbitrary availability patterns, which is what makes the unified many-to-many claim practical.

Reading between the lines

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

  • The fixed split at $T/2$ and the fixed modality ordering for low- and high-frequency selection are never ablated against alternatives, so a natural test is to swap the schedule or move the split point; if performance is insensitive, the gains attributed to FCS would not come from frequency guidance per se.
  • The same coarse-to-fine frequency schedule could transfer to other multi-modal medical synthesis problems, such as CT-to-MRI or PET-to-MRI, where low-frequency anatomy and high-frequency detail play similar roles, though the paper only tests brain MRI.
  • Because the paper measures image-quality metrics rather than downstream task performance, an additional test would be whether synthesized modalities preserve tumor segmentation accuracy or radiomic features; that would tell whether the fidelity gains carry clinical value.
  • A learned or subject-specific frequency-selection strategy might outperform the fixed left-to-right and right-to-left scan, especially across scanners with different contrast properties, but the paper does not explore this direction.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 8 minor

Summary. The paper proposes FgC2F-UDiff, a diffusion-based model for multi-modality MRI synthesis that aims to handle arbitrary combinations of available and missing modalities with a single trained network. It introduces three components: a Coarse-to-fine Unified Network (CUN) that splits the reverse denoising process into coarse and fine stages, a Frequency-guided Collaborative Strategy (FCS) that injects low-frequency guidance early and high-frequency guidance late using a dynamically selected available modality, and a Specific-acceleration Hybrid Mechanism (SHM) combining curriculum learning with stage splitting to reduce sampling steps. Experiments on BraTS 2021 and IXI compare against seven baselines in task-specific and unified settings, reporting PSNR, SSIM, LPIPS, and FID, together with ablations and a time-step analysis.

Significance. If validated, the central claim is practically valuable: one unified model that synthesizes any missing MRI modality from any subset of available modalities, with higher fidelity and fewer denoising steps than current baselines, would address a real clinical bottleneck. The paper's strengths include the breadth of the experimental comparison, the use of two datasets, the release of source code, and the explicit aim of many-to-many synthesis rather than per-task models. However, the key novelty—frequency-staged coarse-to-fine denoising—rests on a self-referential visualization rather than causal ablations, and the reported numerical evidence contains inconsistencies that currently prevent the reader from verifying the superiority claim as stated.

major comments (5)
  1. [V-B vs. V-C, Tables III and V] The same tasks are reported with different PSNR values for 'Our method' in the two comparison tables: T1,T2→T1ce gives 26.62 dB in Table III but 26.34 dB in Table V, and T1,FLAIR→T1ce gives 27.64 dB in Table III but 27.19 dB in Table V. If the Table III results are from a task-specific version of the model while Table V reports the unified model, the manuscript must state this distinction explicitly and explain how the 'superior performance' claim for a unified model is supported by results obtained from a per-task training scheme. As written, the reader cannot tell which variant the abstract and introduction are claiming to be state-of-the-art.
  2. [Table VII, ablation study in V-D] The ablation is internally inconsistent. 'W/O CUN' removes the frequency-guided conditions and drops to 20.58 dB on T1,T2→T1ce, while 'W/O LF or HF' replaces the frequency features with unspecified 'modality-class features' and drops only to 25.35 dB. Both variants remove frequency guidance, so the 4.77 dB gap cannot be attributed to the coarse-to-fine split alone; moreover, the manuscript never defines what 'modality-class features' are or how the network architecture changes between these two ablations. Consequently, the contribution of CUN is not cleanly isolated from the presence or absence of an extra condition channel.
  3. [III-B Discussion and Fig. 3] The load-bearing premise that diffusion denoising progresses from global structure to fine detail, justifying low-frequency guidance in early steps and high-frequency guidance in later steps, is supported only by a visualization of the authors' own model that was trained with exactly this staged LF/HF conditioning. This is self-referential: the observed trajectory may be induced by the conditioning rather than being an independent property that motivates the design. The fixed split at T/2 is never varied, the LF/HF order is never swapped, and a no-staging variant (both frequencies at all steps) is never run. The claimed modality ordering T1 > T2 > FLAIR > T1ce for low-frequency content and its reverse for high-frequency content (Section III-C) is asserted without measurement. Please add these control experiments; without them, the improvements attributed to CUN and FCS are not established.
  4. [IV-C and Tables III-VI] The text states that performance differences were evaluated with signed-rank tests (p < 0.05), but no p-values, test statistics, or multiple-comparison corrections are reported anywhere. In addition, FID is reported without error bars or confidence intervals, although mean ± std is given for PSNR, SSIM, and LPIPS. Please provide the actual significance results and a measure of variation for FID (e.g., across cross-validation folds or multiple sampling runs) to support the claimed statistical superiority.
  5. [V-A and V-C, Tables I-II vs. V-VI] The paper's central many-to-many claim is supported only by Tables I and II, which report the proposed method's absolute results across many input-subsets but contain no baseline comparison for these same many-to-many protocols. The unified comparison in Tables V-VI covers only two many-to-one tasks per dataset. To validate the 'many-to-many superiority' statement, the authors should compare baselines under the same full protocol of all 14 (BraTS) or 6 (IXI) input scenarios, or explicitly restrict the claim to the tasks evaluated.
minor comments (8)
  1. [Eq. (11)] The PSNR formula is written as '10·log10(MAX_I^2√MSE)' which is dimensionally incorrect; the standard form is 10·log10(MAX_I^2 / MSE).
  2. [Eq. (14)] The FID formula writes 'Tr(Σp + Σq − 2(ΣpΣq)1/2)', but the last term should be the trace of the matrix square root of the product ΣpΣq; the notation 'ΣpΣ1/2q' later in the same paragraph is garbled.
  3. [Eq. (1) and surrounding text] The text states 'βt ∼ (0, I)' and calls βt the variance of Gaussian noise, but βt is a scalar schedule and the noise should be ε ∼ N(0, I); please correct the notation to avoid confusion between the variance and the noise term.
  4. [Eq. (7)] The expectation subscript 'Ext,ϵ∼N (0,I),t' is malformed; it should read E_{t,ε∼N(0,I)}.
  5. [Fig. 3 caption and references] The caption 'Visualize analyzing and visualizing the denoising synthesis images' is redundant and should be reworded; references to 'Fig.2.B' and 'Fig.V-D(b)' do not correspond to clearly labeled panels in the figures.
  6. [IV-A and III-C] The frequency-ordering discussion in III-C is framed for four modalities (T1, T2, FLAIR, T1ce), but the IXI dataset has only three (T1, T2, PD); the paper should clarify how FCS generalizes to the IXI modality set and whether the left-to-right/right-to-left ordering is still applied.
  7. [V-E] The term 'over-diffusion' is attributed to reference [71] (DDIM); please verify this citation, as the referenced paper does not appear to introduce that term.
  8. [Tables III-VI] The text repeatedly claims 'p < 0.05' but the tables contain no significance markers; please add p-values or indicate with symbols which comparisons are significant.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: FgC2F-UDiff is an empirical method paper whose claims are evaluated against external baselines and standard datasets, with no equation or prediction reducing by construction to a fit or self-citation.

full rationale

FgC2F-UDiff is an empirical method paper, not a derivation from first principles, so the classic prediction-vs-fit circularity does not arise. The training objective (Eq. 7) is the standard DDPM noise-prediction loss; CUN, FCS, and SHM are architectural and training-strategy choices rather than parameters fitted to the evaluation data and then reported as predictions. All quantitative superiority claims are made against external baselines (pix2pix, pGAN, MM-GAN, ResVit, LDM, CoLa-Diff, Uni-GAN) on public datasets (BraTS 2021 and IXI), so the central claim is externally falsifiable rather than forced by construction. Self-citations (e.g., refs [3], [15], [27], [31]) appear in background and related-work contexts and are not load-bearing; no uniqueness theorem or fitted value from the authors' prior work is invoked to forbid alternatives. The coarse-to-fine premise in Section III-B is motivated by the authors' own visualization (Fig. 3) of their model's denoising trajectory, and the fixed T/2 split is a design constant rather than a fitted parameter; while this is a weakness in evidential support, it does not make the derivation circular because no equation is defined in terms of the claimed outcome and the ablation evidence is not used as a fitted input. The ablation inconsistencies in Table VII are correctness and interpretation concerns, not circularity.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper rests on standard DDPM machinery plus several hand-chosen hyperparameters and unvalidated domain assumptions. The most consequential free parameters are the Gaussian filter width, the fixed coarse-to-fine split at T/2, the number of time steps (T=200), and the hand-defined curriculum difficulty levels. The key ad hoc axioms are the global-to-detail denoising premise and the strict modality ordering for frequency content. No new physical entities are introduced.

free parameters (4)
  • Gaussian filter width sigma (Eq. 8) = not reported
    The frequency decomposition uses a Gaussian kernel with kernel size 21, but the width sigma is not specified anywhere; it controls how much low-frequency versus high-frequency content is separated and is a free choice of the authors.
  • Coarse-to-fine split point = T/2 = 100 (T=200)
    The denoising process is divided at time step T/2. No ablation tests other split points, yet the entire CUN design depends on this boundary.
  • Number of diffusion time steps T = 200
    Chosen to accelerate sampling; the paper shows performance at 200 and 300 steps but does not justify 200 as optimal beyond the figure.
  • Curriculum learning difficulty thresholds = one missing = easy, two = moderate, three = hard
    The ordering of difficulty levels is hand-defined; there is no analysis of the curriculum schedule's sensitivity.
assumptions (5)
  • standard math DDPM forward and reverse processes as formulated in Section III-A are valid for 2D image synthesis.
    The method builds on Ho et al.'s DDPM equations; accepted background.
  • ad hoc to paper Diffusion denoising progresses from global structure to fine detail, so low-frequency guidance is appropriate in early denoising and high-frequency in later denoising.
    Stated in Section III-B and supported only by qualitative visualization in Fig.3; no quantitative or literature-backed evidence, and no ablation of the split.
  • ad hoc to paper The four MRI modalities can be strictly ordered by low-frequency content as T1 > T2 > FLAIR > T1ce and by high-frequency content as T1 < T2 < FLAIR < T1ce, so a left-to-right / right-to-left search finds the best frequency guidance.
    Introduced in Section III-C Discussion; the ordering is asserted from qualitative image characteristics, not measured on the datasets.
  • ad hoc to paper Using one selected available modality's low- or high-frequency image as a condition channel is sufficient and better than fusing frequency information from all available modalities.
    The FCS selects only the first available modality by the ordering; no experiment compares single-modality frequency guidance against multi-modality frequency fusion.
  • domain assumption Synthesizing 2D slices and registering IXI images to T1 space preserves the multi-modal mapping needed for synthesis.
    The method is described for images HxW; volumes are 240x240x150 or 256x256x150, and slice-level processing is assumed without explicit justification.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FgC2F-UDiff: Frequency-guided and Coarse-to-fine Unified Diffusion Model for Multi-modality Missing MRI Synthesis." pith.science (2026). https://pith.science/paper/SASHRVXH

@misc{pith2026250103526,
  author       = {Pith},
  title        = {Pith review of: FgC2F-UDiff: Frequency-guided and Coarse-to-fine Unified Diffusion Model for Multi-modality Missing MRI Synthesis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SASHRVXH}},
  note         = {Machine review of arXiv:2501.03526}
}
read the original abstract

Multi-modality magnetic resonance imaging (MRI) is essential for the diagnosis and treatment of brain tumors. However, missing modalities are commonly observed due to limitations in scan time, scan corruption, artifacts, motion, and contrast agent intolerance. Synthesis of missing MRI has been a means to address the limitations of modality insufficiency in clinical practice and research. However, there are still some challenges, such as poor generalization, inaccurate non-linear mapping, and slow processing speeds. To address the aforementioned issues, we propose a novel unified synthesis model, the Frequency-guided and Coarse-to-fine Unified Diffusion Model (FgC2F-UDiff), designed for multiple inputs and outputs. Specifically, the Coarse-to-fine Unified Network (CUN) fully exploits the iterative denoising properties of diffusion models, from global to detail, by dividing the denoising process into two stages, coarse and fine, to enhance the fidelity of synthesized images. Secondly, the Frequency-guided Collaborative Strategy (FCS) harnesses appropriate frequency information as prior knowledge to guide the learning of a unified, highly non-linear mapping. Thirdly, the Specific-acceleration Hybrid Mechanism (SHM) integrates specific mechanisms to accelerate the diffusion model and enhance the feasibility of many-to-many synthesis. Extensive experimental evaluations have demonstrated that our proposed FgC2F-UDiff model achieves superior performance on two datasets, validated through a comprehensive assessment that includes both qualitative observations and quantitative metrics, such as PSNR SSIM, LPIPS, and FID.

Figures

Figures reproduced from arXiv: 2501.03526 by the authors.

Figure 1
Figure 1. Different image modalities provide different types of tissue contrast [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed FgC2F-UDiff, which cross-modality synthesizes missing modalities from multiple inputs and outputs. It includes forward [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Visualize analyzing and visualizing the denoising synthesis images, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: The image can usually be decomposed into a high-frequency sub [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Fig.5. The four-bit digits in the figures indicate the availability [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 5
Figure 5. Figure 5: Illustrative instances of synthetic images generated by our FgC2F-UDiff on the BraTS Dataset. Each row shows composite diagrams portraying [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Illustrative instances of synthetic images were demonstrated on the BraTS dataset for two representative tasks. Synthesized images from all competing [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Illustrative instances of synthetic images were demonstrated on the IXI dataset for two representative synthesis tasks. Synthesized images from [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Illustrative instances of synthetic images were demonstrated on the [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: The experimental results of our method compared with LDM at [PITH_FULL_IMAGE:figures/full_fig_p012_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

73 extracted references · 50 canonical work pages

  1. [1]

    The multimodal brain tumor image segmentation benchmark (brats),

    B. H. Menze, A. Jakab, S. Bauer, J. Kalpathy-Cramer, K. Farahani, J. Kirby, Y . Burren, N. Porz, J. Slotboom, R. Wiest et al. , “The multimodal brain tumor image segmentation benchmark (brats),” IEEE transactions on medical imaging , vol. 34, no. 10, pp. 1993–2024, 2014

    work page 1993

  2. [2]

    Joint segmentation of anatomical and functional images: Applications in quantification of lesions from pet, pet-ct, mri- pet, and mri-pet-ct images,

    U. Bagci, J. K. Udupa, N. Mendhiratta, B. Foster, Z. Xu, J. Yao, X. Chen, and D. J. Mollura, “Joint segmentation of anatomical and functional images: Applications in quantification of lesions from pet, pet-ct, mri- pet, and mri-pet-ct images,” Medical image analysis , vol. 17, no. 8, pp. 929–945, 2013

    work page 2013

  3. [3]

    Edge-aware multi-task network for integrating quantification segmentation and uncertainty prediction of liver tumor on multi-modality non-contrast mri,

    X. Xiao, Q. V . Hu, and G. Wang, “Edge-aware multi-task network for integrating quantification segmentation and uncertainty prediction of liver tumor on multi-modality non-contrast mri,” inInternational Confer- ence on Medical Image Computing and Computer-Assisted Intervention . Springer, 2023, pp. 652–661

    2023

  4. [4]

    Optimal acquisition sequence for ai-assisted brain tumor segmentation under the constraint of largest information gain per additional mri sequence,

    R. M. Kronberg, D. Meskelevicius, M. Sabel, M. Kollmann, C. Rubbert, and I. Fischer, “Optimal acquisition sequence for ai-assisted brain tumor segmentation under the constraint of largest information gain per additional mri sequence,” Neuroscience Informatics, p. 100053, 2022

    work page 2022

  5. [5]

    Multi- modal mr synthesis via modality-invariant latent representation,

    A. Chartsias, T. Joyce, M. V . Giuffrida, and S. A. Tsaftaris, “Multi- modal mr synthesis via modality-invariant latent representation,” IEEE transactions on medical imaging , vol. 37, no. 3, pp. 803–814, 2017

    work page 2017

  6. [6]

    Pimms: permutation invariant multi-modal segmentation,

    T. Varsavsky, Z. Eaton-Rosen, C. H. Sudre, P. Nachev, and M. J. Cardoso, “Pimms: permutation invariant multi-modal segmentation,” in Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support: 4th International Workshop, DLMIA 2018, and 8th International Workshop, ML-CDS 2018, Held in Conjunction with MICCAI 2018, Granad...

    work page 2018

  7. [7]

    Springer, 2018, pp. 201–209

    work page 2018

  8. [8]

    Image-to-image translation with conditional adversarial networks,

    P. Isola, J.-Y . Zhu, T. Zhou, and A. A. Efros, “Image-to-image translation with conditional adversarial networks,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 1125– 1134

    2017

Show all 73 references
  1. [9]

    Tripartite-gan: Synthesizing liver contrast-enhanced mri to improve tumor detection,

    J. Zhao et al. , “Tripartite-gan: Synthesizing liver contrast-enhanced mri to improve tumor detection,” Medical Image Analysis , vol. 63, p. 101667, 2020

    2020

  2. [10]

    Coca-gan: common-feature-learning-based context-aware generative adversarial network for glioma grading,

    P. Huang, D. Li, Z. Jiao, D. Wei, G. Li, Q. Wang, H. Zhang, and D. Shen, “Coca-gan: common-feature-learning-based context-aware generative adversarial network for glioma grading,” in Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd International Con...

    2019

  3. [11]

    Ea-gans: edge-aware generative adversarial networks for cross-modality mr image synthesis,

    B. Yu, L. Zhou, L. Wang, Y . Shi, J. Fripp, and P. Bourgeat, “Ea-gans: edge-aware generative adversarial networks for cross-modality mr image synthesis,” IEEE transactions on medical imaging , vol. 38, no. 7, pp. 1750–1762, 2019

    2019

  4. [12]

    Unified generative adversarial networks for multimodal segmentation from unpaired 3d medical images,

    W. Yuan, J. Wei, J. Wang, Q. Ma, and T. Tasdizen, “Unified generative adversarial networks for multimodal segmentation from unpaired 3d medical images,” Medical image analysis , vol. 64, p. 101731, 2020

    2020

  5. [13]

    Auto-gan: self- supervised collaborative learning for medical image synthesis,

    B. Cao, H. Zhang, N. Wang, X. Gao, and D. Shen, “Auto-gan: self- supervised collaborative learning for medical image synthesis,” in Pro- ceedings of the AAAI conference on artificial intelligence, vol. 34, no. 07, 2020, pp. 10 486–10 493

    2020

  6. [14]

    Mix-domain contrastive learning for unpaired h&e-to-ihc stain translation,

    S. Wang, Z. Zhang, H. Yan, M. Xu, and G. Wang, “Mix-domain contrastive learning for unpaired h&e-to-ihc stain translation,” arXiv preprint arXiv:2406.11799, 2024

    2024 arXiv

  7. [15]

    mustgan: multi-stream generative adversarial networks for mr image synthesis,

    M. Yurt, S. U. Dar, A. Erdem, E. Erdem, K. K. Oguz, and T. C ¸ ukur, “mustgan: multi-stream generative adversarial networks for mr image synthesis,” Medical image analysis , vol. 70, p. 101944, 2021

    2021

  8. [16]

    Disentangled representation learning for controllable person image generation,

    W. Xu, C. Long, Y . Nie, and G. Wang, “Disentangled representation learning for controllable person image generation,” IEEE Transactions on Multimedia, 2024

    2024

  9. [17]

    Missing mri pulse sequence synthesis using multi-modal generative adversarial network,

    A. Sharma and G. Hamarneh, “Missing mri pulse sequence synthesis using multi-modal generative adversarial network,” IEEE transactions on medical imaging , vol. 39, no. 4, pp. 1170–1183, 2019

    2019

  10. [18]

    Resvit: residual vision transformers for multimodal medical image synthesis,

    O. Dalmaz, M. Yurt, and T. C ¸ ukur, “Resvit: residual vision transformers for multimodal medical image synthesis,” IEEE Transactions on Medical Imaging, vol. 41, no. 10, pp. 2598–2614, 2022

    2022

  11. [19]

    Unified multi-modal image synthesis for missing modality imputation,

    Y . Zhang, C. Peng, Q. Wang, D. Song, K. Li, and S. K. Zhou, “Unified multi-modal image synthesis for missing modality imputation,” arXiv preprint arXiv:2304.05340, 2023

    2023 arXiv

  12. [20]

    Seeing what a gan cannot generate,

    D. Bau, J.-Y . Zhu, J. Wulff, W. Peebles, H. Strobelt, B. Zhou, and A. Torralba, “Seeing what a gan cannot generate,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2019, pp. 4502–4511

    2019

  13. [21]

    Diffusion models beat gans on image synthesis,

    P. Dhariwal and A. Nichol, “Diffusion models beat gans on image synthesis,” Advances in Neural Information Processing Systems, vol. 34, pp. 8780–8794, 2021

    2021

  14. [22]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” Advances in Neural Information Processing Systems , vol. 33, pp. 6840– 6851, 2020

    2020

  15. [23]

    Center-to-edge denoising diffusion probabilistic mod- els with cross-domain attention for undersampled mri reconstruction,

    J. Zhao and S. Li, “Center-to-edge denoising diffusion probabilistic mod- els with cross-domain attention for undersampled mri reconstruction,” in International Conference on Medical Image Computing and Computer- Assisted Intervention. Springer Nature Switzerland, 2024

    2024

  16. [24]

    Srdiff: Single image super-resolution with diffusion probabilistic mod- els,

    H. Li, Y . Yang, M. Chang, S. Chen, H. Feng, Z. Xu, Q. Li, and Y . Chen, “Srdiff: Single image super-resolution with diffusion probabilistic mod- els,” Neurocomputing, vol. 479, pp. 47–59, 2022

    2022

  17. [25]

    Implicit diffusion models for continuous super-resolution,

    S. Gao, X. Liu, B. Zeng, S. Xu, Y . Li, X. Luo, J. Liu, X. Zhen, and B. Zhang, “Implicit diffusion models for continuous super-resolution,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 10 021–10 030

    2023

  18. [26]

    Lfsrdiff: Light field image super-resolution via diffusion models,

    W. Chao, F. Duan, X. Wang, Y . Wang, and G. Wang, “Lfsrdiff: Light field image super-resolution via diffusion models,” arXiv preprint arXiv:2311.16517, 2023

    2023 arXiv

  19. [27]

    Ambiguous medical image segmentation using diffusion models,

    A. Rahman, J. M. J. Valanarasu, I. Hacihaliloglu, and V . M. Patel, “Ambiguous medical image segmentation using diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 11 536–11 546

    2023

  20. [28]

    Multi-layer dense attention decoder for polyp segmentation,

    K. Patel, F. Li, and G. Wang, “Multi-layer dense attention decoder for polyp segmentation,” arXiv preprint arXiv:2403.18180 , 2024

    2024 arXiv

  21. [29]

    United adversarial learning for liver tumor segmenta- tion and detection of multi-modality non-contrast mri,

    J. Zhao et al. , “United adversarial learning for liver tumor segmenta- tion and detection of multi-modality non-contrast mri,” Medical Image Analysis, vol. 73, p. 102154, 2021

    2021

  22. [30]

    Predicting mitral valve mteer surgery outcomes using machine learning and deep learning techniques,

    T. Vyas, M. Chowdhury, X. Xiao, M. Claeys, G. Ong, and G. Wang, “Predicting mitral valve mteer surgery outcomes using machine learning and deep learning techniques,” in Proceedings of the 2024 9th Interna- tional Conference on Mathematics and Artificial Intelligence , 2024, pp. 24–28

    2024

  23. [31]

    Diffusion models for implicit image segmentation ensembles,

    J. Wolleb, R. Sandk ¨uhler, F. Bieder, P. Valmaggia, and P. C. Cattin, “Diffusion models for implicit image segmentation ensembles,” in Inter- national Conference on Medical Imaging with Deep Learning . PMLR, 2022, pp. 1336–1348

    2022

  24. [32]

    Task relevance driven adversarial learning for simultaneous detection, size grading, and quantification of hepato- cellular carcinoma via integrating multi-modality mri,

    X. Xiao, J. Zhao, and S. Li, “Task relevance driven adversarial learning for simultaneous detection, size grading, and quantification of hepato- cellular carcinoma via integrating multi-modality mri,” Medical Image Analysis, vol. 81, p. 102554, 2022

    2022

  25. [33]

    Cola-diff: Conditional latent diffusion model for multi-modal mri synthesis,

    L. Jiang, Y . Mao, X. Wang, X. Chen, and C. Li, “Cola-diff: Conditional latent diffusion model for multi-modal mri synthesis,” in International Conference on Medical Image Computing and Computer-Assisted Inter- vention. Springer, 2023, pp. 398–408. JOURNAL OF LATEX CLASS FILES...

    2023

  26. [34]

    Diffusion models in vision: A survey,

    F.-A. Croitoru, V . Hondru, R. T. Ionescu, and M. Shah, “Diffusion models in vision: A survey,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2023

    2023

  27. [35]

    Conversion between ct and mri images using diffusion and score-matching models,

    Q. Lyu and G. Wang, “Conversion between ct and mri images using diffusion and score-matching models,” arXiv preprint arXiv:2209.12104, 2022

    2022 arXiv

  28. [36]

    Fast patch-based pseudo-ct synthesis from t1-weighted mr images for pet/mr attenuation correction in brain studies,

    A. Torrado-Carvajal, J. L. Herraiz, E. Alcain, A. S. Montemayor, L. Garcia-Canamaque, J. A. Hernandez-Tamames, Y . Rozenholc, and N. Malpica, “Fast patch-based pseudo-ct synthesis from t1-weighted mr images for pet/mr attenuation correction in brain studies,” Journal of Nuclea...

    2016

  29. [37]

    Patch based synthesis of whole head mr images: Application to epi distortion correction,

    S. Roy, Y .-Y . Chou, A. Jog, J. A. Butman, and D. L. Pham, “Patch based synthesis of whole head mr images: Application to epi distortion correction,” in Simulation and Synthesis in Medical Imaging: First Inter- national Workshop, SASHIMI 2016, Held in Conjunction with MICCAI ...

    2016

  30. [38]

    Magnetic resonance image example-based contrast synthesis,

    S. Roy, A. Carass, and J. L. Prince, “Magnetic resonance image example-based contrast synthesis,” IEEE transactions on medical imag- ing, vol. 32, no. 12, pp. 2348–2363, 2013

    2013

  31. [39]

    Pseudo- healthy image synthesis for white matter lesion segmentation,

    C. Bowles, C. Qin, C. Ledig, R. Guerrero, R. Gunn, A. Hammers, E. Sakka, D. A. Dickie, M. V . Hern ´andez, N. Royle et al. , “Pseudo- healthy image synthesis for white matter lesion segmentation,” in Simu- lation and Synthesis in Medical Imaging: First International Workshop, ...

    2016

  32. [40]

    Random forest regression for magnetic resonance image synthesis,

    A. Jog, A. Carass, S. Roy, D. L. Pham, and J. L. Prince, “Random forest regression for magnetic resonance image synthesis,” Medical image analysis, vol. 35, pp. 475–488, 2017

    2017

  33. [41]

    Mr image synthesis by contrast learning on neighborhood en- sembles,

    ——, “Mr image synthesis by contrast learning on neighborhood en- sembles,” Medical image analysis , vol. 24, no. 1, pp. 63–76, 2015

    2015

  34. [42]

    Modality propagation: coherent synthesis of subject-specific scans with data-driven regularization,

    D. H. Ye, D. Zikic, B. Glocker, A. Criminisi, and E. Konukoglu, “Modality propagation: coherent synthesis of subject-specific scans with data-driven regularization,” in Medical Image Computing and Computer-Assisted Intervention–MICCAI 2013: 16th International Con- ference, Nag...

    2013

  35. [43]

    Cross-domain synthesis of medical images using efficient location-sensitive deep network,

    H. Van Nguyen, K. Zhou, and R. Vemulapalli, “Cross-domain synthesis of medical images using efficient location-sensitive deep network,” in Medical Image Computing and Computer-Assisted Intervention– MICCAI 2015: 18th International Conference, Munich, Germany, Oc- tober 5-9, 20...

    2015

  36. [44]

    Whole image synthesis using a deep encoder-decoder network,

    V . Sevetlidis, M. V . Giuffrida, and S. A. Tsaftaris, “Whole image synthesis using a deep encoder-decoder network,” in Simulation and Synthesis in Medical Imaging: First International Workshop, SASHIMI 2016, Held in Conjunction with MICCAI 2016, Athens, Greece, October 21, 20...

    2016

  37. [45]

    Deep learning based imaging data completion for improved brain disease diagnosis,

    R. Li, W. Zhang, H.-I. Suk, L. Wang, J. Li, D. Shen, and S. Ji, “Deep learning based imaging data completion for improved brain disease diagnosis,” in Medical Image Computing and Computer-Assisted Intervention–MICCAI 2014: 17th International Conference, Boston, MA, USA, Septem...

    2014

  38. [46]

    Multi-modality mr image synthesis via confidence-guided aggregation and cross-modality refine- ment,

    B. Peng, B. Liu, Y . Bin, L. Shen, and J. Lei, “Multi-modality mr image synthesis via confidence-guided aggregation and cross-modality refine- ment,” IEEE Journal of Biomedical and Health Informatics , vol. 26, no. 1, pp. 27–35, 2021

    2021

  39. [47]

    Semi-supervised learning of mri synthesis without fully- sampled ground truths,

    M. Yurt, O. Dalmaz, S. Dar, M. Ozbey, B. Tınaz, K. Oguz, and T. C ¸ ukur, “Semi-supervised learning of mri synthesis without fully- sampled ground truths,” IEEE Transactions on Medical Imaging, vol. 41, no. 12, pp. 3895–3906, 2022

    2022

  40. [48]

    Random forest flair reconstruction from t 1, t 2, and p d-weighted mri,

    A. Jog, A. Carass, D. L. Pham, and J. L. Prince, “Random forest flair reconstruction from t 1, t 2, and p d-weighted mri,” in 2014 IEEE 11th International Symposium on Biomedical Imaging (ISBI) . IEEE, 2014, pp. 1079–1082

    2014

  41. [49]

    Rs-net: Regression-segmentation 3d cnn for synthesis of full resolution missing brain mri in the presence of tumours,

    R. Mehta and T. Arbel, “Rs-net: Regression-segmentation 3d cnn for synthesis of full resolution missing brain mri in the presence of tumours,” in Simulation and Synthesis in Medical Imaging: Third International Workshop, SASHIMI 2018, Held in Conjunction with MICCAI 2018, Gran...

    2018

  42. [50]

    Robust multimodal brain tumor segmentation via feature disentanglement and gated fusion,

    C. Chen, Q. Dou, Y . Jin, H. Chen, J. Qin, and P.-A. Heng, “Robust multimodal brain tumor segmentation via feature disentanglement and gated fusion,” in Medical Image Computing and Computer Assisted Intervention–MICCAI 2019: 22nd International Conference, Shenzhen, China, Octo...

    2019

  43. [51]

    Hi-net: hybrid-fusion network for multi-modal mr image synthesis,

    T. Zhou, H. Fu, G. Chen, J. Shen, and L. Shao, “Hi-net: hybrid-fusion network for multi-modal mr image synthesis,” IEEE transactions on medical imaging, vol. 39, no. 9, pp. 2772–2781, 2020

    2020

  44. [52]

    Collagan: Collaborative gan for missing image data imputation,

    D. Lee, J. Kim, W.-J. Moon, and J. C. Ye, “Collagan: Collaborative gan for missing image data imputation,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 2487– 2496

    2019

  45. [53]

    Assessing the importance of magnetic resonance contrasts using collaborative generative adversarial networks,

    D. Lee, W.-J. Moon, and J. C. Ye, “Assessing the importance of magnetic resonance contrasts using collaborative generative adversarial networks,” Nature Machine Intelligence , vol. 2, no. 1, pp. 34–42, 2020

    2020

  46. [54]

    A domain gap aware generative adversarial network for multi-domain image translation,

    W. Xu and G. Wang, “A domain gap aware generative adversarial network for multi-domain image translation,” IEEE Transactions on Image Processing, vol. 31, pp. 72–84, 2021

    2021

  47. [55]

    Multi-modal mri image synthesis via gan with multi-scale gate mergence,

    B. Zhan, D. Li, X. Wu, J. Zhou, and Y . Wang, “Multi-modal mri image synthesis via gan with multi-scale gate mergence,” IEEE Journal of Biomedical and Health Informatics , vol. 26, no. 1, pp. 17–26, 2021

    2021

  48. [56]

    Progressively volumetrized deep generative models for data-efficient contextual learning of mr image recovery,

    M. Yurt, M. ¨Ozbey, S. U. Dar, B. Tinaz, K. K. Oguz, and T. C ¸ ukur, “Progressively volumetrized deep generative models for data-efficient contextual learning of mr image recovery,” Medical Image Analysis , vol. 78, p. 102429, 2022

    2022

  49. [57]

    Rfnet: Region-aware fusion network for incomplete multi-modal brain tumor segmentation,

    Y . Ding, X. Yu, and Y . Yang, “Rfnet: Region-aware fusion network for incomplete multi-modal brain tumor segmentation,” in Proceedings of the IEEE/CVF international conference on computer vision , 2021, pp. 3975–3984

    2021

  50. [58]

    Conditional generative adversarial nets,

    M. Mirza and S. Osindero, “Conditional generative adversarial nets,” arXiv preprint arXiv:1411.1784 , 2014

    2014 arXiv

  51. [59]

    Multi-modal modality- masked diffusion network for brain mri synthesis with random modality missing,

    X. Meng, K. Sun, J. Xu, X. He, and D. Shen, “Multi-modal modality- masked diffusion network for brain mri synthesis with random modality missing,” IEEE Transactions on Medical Imaging , 2024

    2024

  52. [60]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” in Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III

    2015

  53. [61]

    Springer, 2015, pp. 234–241

    2015

  54. [62]

    Gauss and the history of the fast fourier transform,

    M. Heideman, D. Johnson, and C. Burrus, “Gauss and the history of the fast fourier transform,” IEEE Assp Magazine , vol. 1, no. 4, pp. 14–21, 1984

    1984

  55. [63]

    Reconstruction of multispatial, multispectral im- age data using spatial frequency content,

    R. A. Schowengerdt, “Reconstruction of multispatial, multispectral im- age data using spatial frequency content,” Photogrammetric Engineering and Remote Sensing , vol. 46, no. 10, pp. 1325–1334, 1980

    1980

  56. [64]

    Curriculum learning,

    Y . Bengio, J. Louradour, R. Collobert, and J. Weston, “Curriculum learning,” in Proceedings of the 26th annual international conference on machine learning , 2009, pp. 41–48

    2009

  57. [65]

    The rsna-asnr-miccai brats 2021 benchmark on brain tumor segmentation and radiogenomic classification,

    U. Baid, S. Ghodasara, S. Mohan, M. Bilello, E. Calabrese, E. Colak, K. Farahani, J. Kalpathy-Cramer, F. C. Kitamura, S. Pati et al. , “The rsna-asnr-miccai brats 2021 benchmark on brain tumor segmentation and radiogenomic classification,” arXiv preprint arXiv:2107.02314 , 2021

    2021 arXiv

  58. [66]

    Advancing the cancer genome atlas glioma mri collections with expert segmentation labels and radiomic features,

    S. Bakas, H. Akbari, A. Sotiras, M. Bilello, M. Rozycki, J. S. Kirby, J. B. Freymann, K. Farahani, and C. Davatzikos, “Advancing the cancer genome atlas glioma mri collections with expert segmentation labels and radiomic features,” Scientific data, vol. 4, no. 1, pp. 1–13, 2017

    2017

  59. [67]

    Information extraction from images

    IXI. Information extraction from images. [Online]. Available: www. brain-development.org

  60. [68]

    A global optimisation method for robust affine registration of brain images,

    M. Jenkinson and S. Smith, “A global optimisation method for robust affine registration of brain images,”Medical image analysis, vol. 5, no. 2, pp. 143–156, 2001

    2001

  61. [69]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980 , 2014

    2014 arXiv

  62. [70]

    Image quality assessment: from error visibility to structural similarity,

    Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: from error visibility to structural similarity,” IEEE transactions on image processing , vol. 13, no. 4, pp. 600–612, 2004

    2004

  63. [71]

    Image synthesis in multi-contrast mri with conditional generative adversarial networks,

    S. U. Dar, M. Yurt, L. Karacan, A. Erdem, E. Erdem, and T. Cukur, “Image synthesis in multi-contrast mri with conditional generative adversarial networks,” IEEE transactions on medical imaging , vol. 38, no. 10, pp. 2375–2388, 2019

    2019

  64. [72]

    High- resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High- resolution image synthesis with latent diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2022, pp. 10 684–10 695

    2022

  65. [73]

    Denoising diffusion implicit models,

    J. Song, C. Meng, and S. Ermon, “Denoising diffusion implicit models,” arXiv preprint arXiv:2010.02502 , 2020

    2010 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.