Pith. sign in

REVIEW 4 major objections 6 minor 2 cited by

ContextMRI: Enhancing Compressed Sensing MRI through Metadata Conditioning

T0 review · 4 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read Conditioning MRI reconstruction on clinical metadata consistently improves accuracy.

desk verdict Plausible and useful engineering contribution, but the headline claim of consistent gains is not statistically secured by the reported evidence. read the letter →

arxiv 2501.04284 v2 pith:FP3RTX43 submitted 2025-01-08 cs.CV cs.LG

classification cs.CVcs.LG
keywords ContextMRIcompressedsensingdiffusionmodelsmetadataconditioningtext-conditionedgenerationinverseproblemsreconstructionclassifier-freeguidance
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 argues that clinically recorded metadata—anatomy, slice location, contrast, MR pulse-sequence parameters, patient age and sex, and pathology—carries information that current compressed-sensing MRI reconstructions throw away. It proposes ContextMRI, a text-conditioned diffusion model trained on complex-valued MRI images, and uses it as a prior in a diffusion-based inverse solver. The central claim is that feeding this metadata into the prior yields more accurate reconstructions than an unconditional prior, with gains that persist across datasets, acceleration factors, and undersampling patterns, and that the gains grow as the metadata becomes more complete. If true, this means already-available clinical records can improve image quality at no extra scan time.

What carries the argument

The load-bearing object is a text-conditioned pixel-space diffusion model trained directly on complex-valued MR images represented as two channels, one real and one imaginary. Metadata is flattened into a structured sentence such as “Knee, Slice 19, PDFS, Pathology: ..., TR: 3150, TE: 33, TI: 100, Flip angle: 150”; a frozen contrastive text encoder turns this sentence into an embedding that modulates the diffusion U-Net. At inference, the conditional prior is combined with the measurement model by a decomposed diffusion sampler: each denoising step predicts the clean image, then solves a small data-consistency optimization problem with conjugate-gradient steps, and classifier-free guidance is used to accentuate the metadata signal. This design lets the same pretrained prior be reused zero-shot for different masks and acceleration factors, and it is what makes conditioning on metadata a concrete, testable intervention rather than a vague qualitative advantage.

What would settle it

Reconstruct the same set of slices from the same undersampled measurements many times with different random seeds, once with metadata conditioning and once without, and compute paired per-slice PSNR/SSIM differences. If the 95% confidence interval for the mean difference includes zero in the uniform-1D ×4 setting, the paper's central claim of consistent gains would not be supported; if it excludes zero, the claim survives.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that a diffusion prior conditioned on granular clinical metadata solves compressed-sensing MRI better than the same prior without conditioning. The authors train a pixel-space diffusion model on minimum-variance-unbiased complex-valued images, convert metadata into a structured text prompt, and encode it with a frozen contrastive text encoder; at inference they plug this conditional prior into a decomposed diffusion sampler that alternates denoising steps with conjugate-gradient data-consistency steps, using classifier-free guidance to control how strongly the metadata steers the result. Across uniform one-dimensional and Poisson two-dimensional undersampling masks at ×4, ×8, and ×15 acceleration, conditional reconstruction outperforms the unconditional baseline, and removing metadata fields one at a time—MR parameters, contrast, slice, demographics, pathology—degrades performance in steps. The authors also report that supplying a wrong pathology label still beats no label, because it signals the presence of an abnormality.

Load-bearing premise

The claim that conditional reconstruction consistently beats unconditional reconstruction assumes that the observed mean PSNR/SSIM differences, which are small relative to the reported slice-to-slice standard deviations, are real improvements rather than sampling noise.

Editorial extensions

If this is right

  • Metadata already recorded during clinical scans becomes a free source of reconstruction signal: no additional acquisition time or hardware is needed.
  • Reconstruction quality improves stepwise as more metadata fields are provided, so hospitals with richer structured records would see larger benefits.
  • Even a wrong pathology impression helps relative to no conditioning, as long as it indicates that some abnormality is present, which could make the method robust to preliminary clinical notes.
  • Conditional reconstruction tends to reduce slice-to-slice variability, so the same prior could make accelerated scans more consistent for downstream analysis.

Reading between the lines

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

  • A paired significance analysis of per-slice conditional-versus-unconditional differences would directly test whether the reported 0.2–0.5 dB gains are above noise; the reported standard deviations make this the obvious next check.
  • The same conditioning recipe could transfer to other medical inverse problems—CT, PET, ultrasound—where acquisition parameters and demographics are already stored in standard image headers.
  • If the gains hold up, metadata should be treated as part of the forward model rather than an optional prompt: the reconstruction posterior becomes p(x | y, c) instead of p(x | y), a shift that touches every learned inverse solver, not just MRI.
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

4 major / 6 minor

Summary. The paper proposes ContextMRI, a pixel-space text-conditioned diffusion model for complex-valued MRI. The model is trained on MVUE complex images from fastMRI knee and brain, with metadata (anatomy, slice, contrast, MR parameters, pathology) formatted as structured text prompts and encoded with a frozen CLIP text encoder. At inference, the conditional prior is combined with the Decomposed Diffusion Sampler (DDS) for compressed-sensing MRI reconstruction. The central claims are that metadata conditioning yields consistent reconstruction gains across datasets, acceleration factors, and undersampling patterns, and that increasing metadata fidelity systematically improves performance. Additional experiments on SKM-TEA address patient demographics, and fastMRI+ pathology annotations are used to test robustness to incorrect pathology labels.

Significance. If the claims hold, this is a useful and timely direction: clinically available metadata can be exploited at zero acquisition cost to condition MRI reconstruction priors. The paper has several concrete strengths: it handles complex-valued images directly, trains at scale (167,375 slices for fastMRI; 25,731 for SKM-TEA), releases code, and uses a robust inverse-problem solver (DDS). The idea of treating metadata as structured text and using CLIP embeddings is simple and reproducible. However, the significance is currently tempered by the statistical support: the reported effect sizes are small relative to the reported variability, several Table 1 entries contradict the 'consistent gains' claim, and the figures supporting the metadata-fidelity ordering lack error bars. The engineering contribution is solid, but the abstract's strong claims need to be reined in or backed by proper statistical analysis.

major comments (4)
  1. [Section 2.2, Table 1] The claim of 'consistent gains regardless of the circumstances' is contradicted by the paper's own table. For knee Poisson2D ×8 at CFG=3, PSNR is 33.98±5.85 versus 34.45±4.65 for the unconditional baseline, and LPIPS worsens from 12.46 to 14.22. For brain uniform1D ×8 at CFG=3, LPIPS worsens from 19.77 to 20.97. Several other CFG=3 entries also degrade relative to CFG=1 or CFG=2. These reversals undermine the abstract's 'consistent gains across multiple datasets, acceleration factors, and undersampling patterns' and Section 2.2's 'regardless of anatomy, mask type, or sampling ratio'.
  2. [Section 2.2, Table 1] No statistical significance testing is reported. The mean PSNR gains are roughly 0.1–0.5 dB, while the reported standard deviations are 2–6 dB, so the effect sizes are small relative to the spread. Moreover, slices from the same volume are highly correlated, so a naive per-slice analysis would overstate significance; the analysis should cluster by volume or by patient. Without paired tests, confidence intervals, or clustered error bars, the statement that ContextMRI 'significantly surpasses' the unconditional DDS baseline is not statistically supported.
  3. [Section 2.3, Figs. 3–4] The claim that removing MR parameters, contrast, or slice information produces a 'significant gap' and a 'gradual decrease' in performance rests on figures whose y-axes span sub-decibel ranges (e.g., Fig. 3: 30.5–31.3 dB; Fig. 4: 27.78–27.86 dB) and which show no error bars or confidence intervals. The same issue applies to the patient-demographic comparison in Fig. 4. The paper should provide error bars or confidence intervals and a formal test of the monotonic ordering of metadata fidelity, not just visual inspection of zoomed curves.
  4. [Section 2.5, Fig. 5] The robustness claim that using a wrong pathology label still yields reconstructions 'better than the unconditional case' is based on curves without error bars or quantitative tabulation. Given the small effect sizes and reversals documented in Table 1, the paper should report the actual PSNR/SSIM/LPIPS values and their variability for the correct-pathology, wrong-pathology, and unconditional conditions, and ideally a paired test, before claiming that incorrect pathology provides a 'net boost'.
minor comments (6)
  1. [Section 2.1] The text refers to 'Figure ??' twice in the discussion of metadata-conditioned synthesis; these placeholders must be resolved before publication.
  2. [Section 1] There is a typo in the first paragraph: 'appcliations' should be 'applications' (in the sentence about heuristic tuning of algorithms).
  3. [Section 4.4] The spacing in 'V AE' should be corrected to 'VAE'; also 'T5-XXL' is inconsistently written as 'T5-XXL' in some places and 'T5-XXL' in others.
  4. [Table 2] The sequence names 'Turbospinecho' and 'QDess' should be standardized to conventional terminology such as 'Turbo Spin Echo' and 'QDESS' for clarity.
  5. [Section 2.2] The sentence 'This indicates that metadata injection augments the model’s prior beyond information already contained in the undersampled measurement' is a causal interpretation that is not directly tested; alternative explanations, such as CFG acting as a regularizer or changing the effective sampling temperature, are not considered.
  6. [Section 1] The data processing inequality reference [11] is used rhetorically; the statement that ignoring metadata is 'a suboptimal practice according to the data processing inequality' is not a formal argument, since conditioning on c is not a post-processing operation on y alone. Consider softening or clarifying this point.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the metadata-conditioning claim is an empirical result, not a derivation, and the DDS self-citation is a baseline rather than a load-bearing premise.

full rationale

ContextMRI's central claim is that conditioning a diffusion prior on clinical metadata improves CS-MRI reconstruction. This is established by training a pixel-space diffusion model and comparing measured reconstruction metrics under different CFG scales, metadata ablations, datasets, acceleration factors, and undersampling patterns. No equation in Section 4 derives the reported PSNR/SSIM/LPIPS gains from the definition of the method; the gains are measured outcomes, not consequences of a fitted parameter renamed as a prediction. The data processing inequality is invoked rhetorically in the introduction and in Section 2.3 ('highlighting the advantage of ContextMRI arising from the data processing inequality'), not as a derivation, and the paper does not claim to prove improvement from the inequality. The diffusion sampler DDS [7] is prior work by the same authors and is used as the inference backbone, but it is also the baseline against which the conditional and unconditional versions are compared; the metadata benefit does not reduce to the DDS equations (Eqs. 4-6). There is no uniqueness theorem imported from the authors, no ansatz smuggled in via citation, and no known empirical result renamed as organization. The main weakness is statistical: reported gains are roughly 0.1-0.5 dB with standard deviations of 2-6 dB, several Table 1 entries reverse direction, and no significance tests or volume-clustered error bars are provided. That is a correctness and robustness concern, not circularity. Since no circular step can be quoted and exhibited, the appropriate score under the proportional scale is 0.

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

The paper introduces no new physical entities. Its load-bearing assumptions are domain-level: CLIP embeddings work for MRI metadata, ESPIRiT/MVUE preprocessing yields a trustworthy training target, the forward model is accurate, and pathology/demographic labels are reliable. The only tuned numerical parameters are the CFG scale, DDS hyperparameters, and a training drop probability; none are fitted to the test set, but they are hand-chosen.

free parameters (5)
  • CFG scale gamma = 0.0 to 3.0, swept per experiment
    Conditioning strength is tuned by evaluating multiple values; the claim of improvement holds for some but not all gamma values, and performance drops beyond 3.0.
  • DDS hyperparameter eta = 0.8
    Hand-chosen constant controlling stochasticity in the DDIM sampling step of the inverse solver.
  • DDS data consistency weight xi = 5.0
    Hand-chosen weight balancing data fidelity against the prior in the optimization step of DDS.
  • Number of conjugate gradient steps in DDS = 5
    Hand-chosen number of solver iterations; affects reconstruction quality and is not derived from first principles.
  • Random drop probability for MR imaging parameters = 0.5
    During training, TR, TE, TI, and flip angle are dropped with 50% probability to improve robustness; this value is chosen by hand.
assumptions (5)
  • domain assumption CLIP text embeddings of structured metadata prompts capture enough semantic information to condition MRI generation and reconstruction.
    The frozen CLIP encoder (Section 4.4) was trained on natural images and text; the authors note it may not understand MRI-specific terminology (Discussion), yet the entire method depends on its embeddings being informative.
  • domain assumption ESPIRiT sensitivity maps and MVUE estimates provide a valid ground truth for training the diffusion model.
    Section 4.1 uses ESPIRiT to estimate sensitivity maps and computes MVUE images; if these are biased or inaccurate, the learned prior is imperfect, though this is a standard practice.
  • standard math The forward model y = Ax + n with known coil sensitivity maps accurately represents the MRI acquisition.
    Section 4.5 defines the multi-coil forward model and uses it in the data consistency step; this is a standard linear model but assumes known sensitivities and Gaussian noise.
  • domain assumption fastMRI+ pathology annotations, with bounding box locations ignored, are sufficiently accurate and informative for conditioning.
    Section 4.2 uses pathology labels from fastMRI+ and excludes annotations below a confidence threshold for SKM-TEA, but the reliability of these labels for training is assumed.
  • domain assumption At inference time, the required metadata (slice, contrast, MR parameters, and where applicable demographics and pathology) is known and can be faithfully encoded.
    The method assumes the metadata used to condition the model is available at test time, which may not hold in all clinical settings; the authors discuss this heterogeneity (Discussion).

how reviews work

0 comments
Cite this review

Pith. "Pith review of ContextMRI: Enhancing Compressed Sensing MRI through Metadata Conditioning." pith.science (2026). https://pith.science/paper/FP3RTX43

@misc{pith2026250104284,
  author       = {Pith},
  title        = {Pith review of: ContextMRI: Enhancing Compressed Sensing MRI through Metadata Conditioning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FP3RTX43}},
  note         = {Machine review of arXiv:2501.04284}
}
read the original abstract

Compressed sensing MRI seeks to accelerate MRI acquisition processes by sampling fewer k-space measurements and then reconstructing the missing data algorithmically. The success of these approaches often relies on strong priors or learned statistical models. While recent diffusion model-based priors have shown great potential, previous methods typically ignore clinically available metadata (e.g. patient demographics, imaging parameters, slice-specific information). In practice, metadata contains meaningful cues about the anatomy and acquisition protocol, suggesting it could further constrain the reconstruction problem. In this work, we propose ContextMRI, a text-conditioned diffusion model for MRI that integrates granular metadata into the reconstruction process. We train a pixel-space diffusion model directly on minimally processed, complex-valued MRI images. During inference, metadata is converted into a structured text prompt and fed to the model via CLIP text embeddings. By conditioning the prior on metadata, we unlock more accurate reconstructions and show consistent gains across multiple datasets, acceleration factors, and undersampling patterns. Our experiments demonstrate that increasing the fidelity of metadata, ranging from slice location and contrast to patient age, sex, and pathology, systematically boosts reconstruction performance. This work highlights the untapped potential of leveraging clinical context for inverse problems and opens a new direction for metadata-driven MRI reconstruction.

Figures

Figures reproduced from arXiv: 2501.04284 by the authors.

Figure 1
Figure 1. Illustration of the method used in ContextMRI. (a) We convert available metadata into text format, which is encoded as a feature vector as an additional input to the diffusion model. The diffusion model is trained in pixel space with MVUE complex-valued images. (b) ContextMRI can be used for CS-MRI by leveraging off-the-shelf diffusion model-based inverse problem solvers while additionally incorporating available me… view at source ↗
Figure 2
Figure 2. Quantitative and qualitative comparison of the proposed method against unconditional recon￾struction (DDS). The dashed line with † indicates that DDS was performed with the diffusion model used in the original work of [7], which is the reason for the difference in performance between the unconditional version of our method and the dotted line. ditioning. Roentgen [1] finetunes Stable Diffusion for X-ray synthesis fr… view at source ↗
Figure 3
Figure 3. PSNR vs. CFG by varying the amount of information contained in the metadata. Experiments are conducted on a subset of fastMRI knee data with uniform 1D ×4 acceleration. It may be clear from the image syn￾thesis experiment that the obviously human-interpretable metadata includ￾ing the anatomy, slice location, and contrast will have a significant im￾pact on the image reconstruction pro￾cedure. In contrast, the differe… view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: PSNR vs. CFG on SKM-TEA validation set with and without patient demographic (i.e. age, sex) information. In essence, these parameters represent the “recipe” of the pulse sequence that determines how different tissue com￾ponents appear in the final image. By encoding th…
Figure 5
Figure 5. Figure 5: Trend in PSNR vs. CFG when the pathology of the scan is known. The first row shows 4 different pathologies that are the most dominant in the fastMRI knee dataset. The second row shows 4 different pathologies in the fastMRI brain dataset. All experiments are done with u…
Figure 6
Figure 6. Figure 6: Metadata statistics of fastMRI knee train set 14 [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]
Figure 7
Figure 7. Figure 7: Metadata statistics of fastMRI brain train set 15 [PITH_FULL_IMAGE:figures/full_fig_p015_7.png]
Figure 8
Figure 8. Figure 8: Metadata statistics of SKM-TEA dataset 16 [PITH_FULL_IMAGE:figures/full_fig_p016_8.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Inference-Time Search Using Side Information for Diffusion-Based Image Reconstruction

    cs.CV 2025-10 conditional novelty 6.0 of 10

    Injecting side information via inference-time particle search (GS/RFJS) improves diffusion-based inverse problem reconstructions across inpainting, super-resolution, deblurring, and MRI tasks in a training-free, plug-...

  2. MR-CLIP: Efficient Metadata-Guided Learning of MRI Contrast Representations

    cs.CV 2025-06 conditional novelty 6.0 of 10

    MR-CLIP aligns MRI slices with DICOM acquisition metadata through supervised contrastive learning, yielding contrast-aware representations that transfer to an unseen dataset.

Reference graph

Works this paper leans on

49 extracted references · 35 canonical work pages · cited by 2 Pith papers

  1. [1]

    A vision–language foundation model for the generation of realistic chest x-ray images

    Christian Bluethgen, Pierre Chambon, Jean-Benoit Delbrouck, Rogier van der Sluijs, Mał- gorzata Połacin, Juan Manuel Zambrano Chaves, Tanishq Mathew Abraham, Shivanshu Purohit, Curtis P Langlotz, and Akshay S Chaudhari. “A vision–language foundation model for the generation of realistic chest x-ray images”. In: Nature Biomedical Engineering (2024), pp. 1–...

  2. [2]

    Classifier-Free Guidance is a Predictor-Corrector

    Arwen Bradley and Preetum Nakkiran. “Classifier-Free Guidance is a Predictor-Corrector”. In: arXiv preprint arXiv:2408.09000 (2024) (page 4)

  3. [3]

    Video generation models as world simulators

    Tim Brooks, Bill Peebles, Connor Holmes, Will DePue, Yufei Guo, Li Jing, David Schnurr, Joe Taylor, Troy Luhman, Eric Luhman, Clarence Ng, Ricky Wang, and Aditya Ramesh. “Video generation models as world simulators”. In: (2024). URL: https://openai.com/ research/video-generation-models-as-world-simulators (page 1)

  4. [4]

    MediSyn: Text-Guided Diffusion Models for Broad Medical 2D and 3D Image Synthesis

    Joseph Cho, Cyril Zakka, Rohan Shad, Ross Wightman, Akshay Chaudhari, and William Hiesinger. “MediSyn: Text-Guided Diffusion Models for Broad Medical 2D and 3D Image Synthesis”. In: arXiv preprint arXiv:2405.09806 (2024) (page 3)

  5. [5]

    Diffusion Posterior Sampling for General Noisy Inverse Problems

    Hyungjin Chung, Jeongsol Kim, Michael Thompson Mccann, Marc Louis Klasky, and Jong Chul Ye. “Diffusion Posterior Sampling for General Noisy Inverse Problems”. In:International Conference on Learning Representations. 2023. URL: https://openreview.net/forum? id=OnD9zGAGT0k (pages 1, 7)

  6. [6]

    Cfg++: Manifold-constrained classifier free guidance for diffusion models

    Hyungjin Chung, Jeongsol Kim, Geon Yeong Park, Hyelin Nam, and Jong Chul Ye. “Cfg++: Manifold-constrained classifier free guidance for diffusion models”. In: arXiv preprint arXiv:2406.08070 (2024) (page 4)

  7. [7]

    Decomposed Diffusion Sampler for Accel- erating Large-Scale Inverse Problems

    Hyungjin Chung, Suhyeon Lee, and Jong Chul Ye. “Decomposed Diffusion Sampler for Accel- erating Large-Scale Inverse Problems”. In: The Twelfth International Conference on Learning Representations. 2024. URL: https://openreview.net/forum?id=DsEhqQtfAG (pages 3, 5, 9)

  8. [8]

    Score-based diffusion models for accelerated MRI

    Hyungjin Chung and Jong Chul Ye. “Score-based diffusion models for accelerated MRI”. In: Medical Image Analysis (2022), p. 102479 (page 1)

Show all 49 references
  1. [9]

    Prompt-tuning Latent Diffusion Models for Inverse Problems

    Hyungjin Chung, Jong Chul Ye, Peyman Milanfar, and Mauricio Delbracio. “Prompt-tuning Latent Diffusion Models for Inverse Problems”. In: Forty-first International Conference on Machine Learning. 2024. URL: https : / / openreview . net / forum ? id = hrwIndai8e (pages 1, 2, 9)

  2. [10]

    Lu- ciddreamer: Domain-free generation of 3d gaussian splatting scenes

    Jaeyoung Chung, Suyoung Lee, Hyeongjin Nam, Jaerin Lee, and Kyoung Mu Lee. “Lu- ciddreamer: Domain-free generation of 3d gaussian splatting scenes”. In: arXiv preprint arXiv:2311.13384 (2023) (page 7). 10

  3. [11]

    Elements of information theory

    Thomas M Cover. Elements of information theory. John Wiley & Sons, 1999 (page 2)

  4. [12]

    A survey on diffusion models for inverse problems

    Giannis Daras, Hyungjin Chung, Chieh-Hsin Lai, Yuki Mitsufuji, Jong Chul Ye, Peyman Milanfar, Alexandros G Dimakis, and Mauricio Delbracio. “A survey on diffusion models for inverse problems”. In: arXiv preprint arXiv:2410.00083 (2024) (page 9)

  5. [13]

    SKM-TEA: A Dataset for Accelerated MRI Reconstruction with Dense Image Labels for Quantitative Clinical Evaluation

    Arjun D Desai, Andrew M Schmidt, Elka B Rubin, Christopher Michael Sandino, Marianne Susan Black, Valentina Mazzoli, Kathryn J Stevens, Robert Boutin, Christopher Re, Garry E Gold, et al. “SKM-TEA: A Dataset for Accelerated MRI Reconstruction with Dense Image Labels for Quanti...

  6. [14]

    Parallel MR imaging

    Anagha Deshmane, Vikas Gulani, Mark A Griswold, and Nicole Seiberlich. “Parallel MR imaging”. In: Journal of Magnetic Resonance Imaging 36.1 (2012), pp. 55–72 (page 9)

  7. [15]

    Novel View Synthesis with Pixel-Space Diffusion Models

    Noam Elata, Bahjat Kawar, Yaron Ostrovsky-Berman, Miriam Farber, and Ron Sokolovsky. “Novel View Synthesis with Pixel-Space Diffusion Models”. In:arXiv preprint arXiv:2411.07765 (2024) (page 9)

  8. [16]

    GenerateCT: Text-Conditional Generation of 3D Chest CT V olumes

    Ibrahim Ethem Hamamci, Sezgin Er, Anjany Sekuboyina, Enis Simsar, Alperen Tezcan, Ayse Gulnihan Simsek, Sevval Nil Esirgun, Furkan Almas, Irem Dogan, Muhammed Furkan Dasdelen, et al. “GenerateCT: Text-Conditional Generation of 3D Chest CT V olumes”. In: arXiv preprint arXiv:23...

  9. [17]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. “Denoising diffusion probabilistic models”. In: Advances in Neural Information Processing Systems 33 (2020), pp. 6840–6851 (pages 1, 8, 9)

  10. [18]

    Classifier-Free Diffusion Guidance

    Jonathan Ho and Tim Salimans. “Classifier-Free Diffusion Guidance”. In: NeurIPS 2021 Workshop on Deep Generative Models and Downstream Applications . 2021. URL: https: //openreview.net/forum?id=qw8AKxfYbI (pages 1, 8)

  11. [19]

    Estimation of non-normalized statistical models by score matching

    Aapo Hyvärinen and Peter Dayan. “Estimation of non-normalized statistical models by score matching.” In: Journal of Machine Learning Research 6.4 (2005) (page 8)

  12. [20]

    Robust compressed sensing mri with deep generative priors

    Ajil Jalal, Marius Arvinte, Giannis Daras, Eric Price, Alexandros G Dimakis, and Jonathan Tamir. “Robust compressed sensing mri with deep generative priors”. In:Advances in Neural Information Processing Systems 34 (2021) (pages 1, 4)

  13. [21]

    Stochastic Solutions for Linear Inverse Problems using the Prior Implicit in a Denoiser

    Zahra Kadkhodaie and Eero P Simoncelli. “Stochastic Solutions for Linear Inverse Problems using the Prior Implicit in a Denoiser”. In: Advances in Neural Information Processing Systems. Ed. by A. Beygelzimer, Y . Dauphin, P. Liang, and J. Wortman Vaughan. 2021.URL: https://ope...

  14. [22]

    Denoising Diffusion Restora- tion Models

    Bahjat Kawar, Michael Elad, Stefano Ermon, and Jiaming Song. “Denoising Diffusion Restora- tion Models”. In: Advances in Neural Information Processing Systems. Ed. by Alice H. Oh, Alekh Agarwal, Danielle Belgrave, and Kyunghyun Cho. 2022. URL: https://openreview. net/forum?id=...

  15. [23]

    Denoising diffusion probabilistic models for 3D medical image generation

    Firas Khader, Gustav Müller-Franzes, Soroosh Tayebi Arasteh, Tianyu Han, Christoph Haar- burger, Maximilian Schulze-Hagen, Philipp Schad, Sandy Engelhardt, Bettina Baeßler, Sebas- tian Foersch, et al. “Denoising diffusion probabilistic models for 3D medical image generation”. ...

  16. [24]

    Regularization by Texts for Latent Diffusion Inverse Solvers

    Jeongsol Kim, Geon Yeong Park, Hyungjin Chung, and Jong Chul Ye. “Regularization by Texts for Latent Diffusion Inverse Solvers”. In: arXiv preprint arXiv:2311.15658 (2023) (pages 1, 2, 9)

  17. [25]

    Dreamsampler: Unifying diffusion sampling and score distillation for image manipulation

    Jeongsol Kim, Geon Yeong Park, and Jong Chul Ye. “Dreamsampler: Unifying diffusion sampling and score distillation for image manipulation”. In:European Conference on Computer Vision. Springer. 2025, pp. 398–414 (pages 1, 2, 9)

  18. [26]

    Controllable text-to-image synthesis for multi-modality MR images

    Kyuri Kim, Yoonho Na, Sung-Joon Ye, Jimin Lee, Sung Soo Ahn, Ji Eun Park, and Hwiyoung Kim. “Controllable text-to-image synthesis for multi-modality MR images”. In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2024, pp. 7936–7945 (page 3)

  19. [27]

    Auto-encoding variational bayes

    Diederik P Kingma and Max Welling. “Auto-encoding variational bayes”. In: arXiv preprint arXiv:1312.6114 (2013) (page 9)

  20. [28]

    The multimodal brain tumor image segmentation benchmark (BRATS)

    Bjoern H Menze, Andras Jakab, Stefan Bauer, Jayashree Kalpathy-Cramer, Keyvan Farahani, Justin Kirby, Yuliya Burren, Nicole Porz, Johannes Slotboom, Roland Wiest, et al. “The multimodal brain tumor image segmentation benchmark (BRATS)”. In: IEEE transactions on medical imaging...

  21. [29]

    SigPy: a python package for high performance iterative reconstruction

    Frank Ong and Michael Lustig. “SigPy: a python package for high performance iterative reconstruction”. In: Proceedings of the ISMRM 27th Annual Meeting, Montreal, Quebec, Canada. V ol. 4819. 5. 2019 (page 10)

  22. [30]

    Scalable diffusion models with transformers

    William Peebles and Saining Xie. “Scalable diffusion models with transformers”. In: Proceed- ings of the IEEE/CVF International Conference on Computer Vision. 2023, pp. 4195–4205 (page 1)

  23. [31]

    Würstchen: An efficient architecture for large-scale text-to-image diffusion models

    Pablo Pernias, Dominic Rampas, Mats L Richter, Christopher J Pal, and Marc Aubreville. “Würstchen: An efficient architecture for large-scale text-to-image diffusion models”. In:arXiv preprint arXiv:2306.00637 (2023) (page 3)

  24. [32]

    Brain imaging generation with latent diffusion models

    Walter HL Pinaya, Petru-Daniel Tudosiu, Jessica Dafflon, Pedro F Da Costa, Virginia Fernan- dez, Parashkev Nachev, Sebastien Ourselin, and M Jorge Cardoso. “Brain imaging generation with latent diffusion models”. In: MICCAI Workshop on Deep Generative Models. Springer. 2022, p...

  25. [33]

    DreamFusion: Text-to-3D using 2D Diffusion

    Ben Poole, Ajay Jain, Jonathan T. Barron, and Ben Mildenhall. “DreamFusion: Text-to-3D using 2D Diffusion”. In: arXiv (2022) (page 4)

  26. [34]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. “Learning transferable visual models from natural language supervision”. In: International conference on machine learning....

  27. [35]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. “High-resolution image synthesis with latent diffusion models”. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022, pp. 10684–10695 (pages 1, 8, 9)

  28. [36]

    Solving linear inverse problems provably via posterior sampling with latent diffusion models

    Litu Rout, Negin Raoof, Giannis Daras, Constantine Caramanis, Alex Dimakis, and Sanjay Shakkottai. “Solving linear inverse problems provably via posterior sampling with latent diffusion models”. In: Advances in Neural Information Processing Systems 36 (2024) (page 9)

  29. [37]

    Photorealistic text-to-image diffusion models with deep language understanding

    Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L Denton, Kamyar Ghasemipour, Raphael Gontijo Lopes, Burcu Karagol Ayan, Tim Salimans, et al. “Photorealistic text-to-image diffusion models with deep language understanding”. In: Ad- vances in neural inf...

  30. [38]

    Diffusion Models as Data Mining Tools

    Ioannis Siglidis, Aleksander Holynski, Alexei A Efros, Mathieu Aubry, and Shiry Ginosar. “Diffusion Models as Data Mining Tools”. In:arXiv preprint arXiv:2408.02752(2024) (page 7)

  31. [39]

    Denoising Diffusion Implicit Models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. “Denoising Diffusion Implicit Models”. In: 9th International Conference on Learning Representations, ICLR. 2021 (page 9)

  32. [40]

    Solving Inverse Problems in Medical Imaging with Score-Based Generative Models

    Yang Song, Liyue Shen, Lei Xing, and Stefano Ermon. “Solving Inverse Problems in Medical Imaging with Score-Based Generative Models”. In:International Conference on Learning Rep- resentations. 2022. URL: https://openreview.net/forum?id=vaRCHVj0uGI (page 1)

  33. [41]

    Score-Based Generative Modeling through Stochastic Differential Equations

    Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. “Score-Based Generative Modeling through Stochastic Differential Equations”. In: 9th International Conference on Learning Representations, ICLR. 2021 (pages 1, 8)

  34. [42]

    UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age

    Cathie Sudlow, John Gallacher, Naomi Allen, Valerie Beral, Paul Burton, John Danesh, Paul Downey, Paul Elliott, Jane Green, Martin Landray, et al. “UK biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age”. In: PL...

  35. [43]

    DeepFloyd-IF: High-Quality Text-to-Image Synthesis

    DeepFloyd Team. DeepFloyd-IF: High-Quality Text-to-Image Synthesis. https://github. com/deepfloyd/IF. Accessed: 2024-09-12. 2023 (pages 8, 9)

  36. [44]

    ESPIRiT—an eigenvalue approach to autocalibrating parallel MRI: where SENSE meets GRAPPA

    Martin Uecker, Peng Lai, Mark J Murphy, Patrick Virtue, Michael Elad, John M Pauly, Shreyas S Vasanawala, and Michael Lustig. “ESPIRiT—an eigenvalue approach to autocalibrating parallel MRI: where SENSE meets GRAPPA”. In: Magnetic resonance in medicine 71.3 (2014), pp. 990–100...

  37. [45]

    Towards General Text-guided Image Synthesis for Customized Multimodal Brain MRI Generation

    Yulin Wang, Honglin Xiong, Kaicong Sun, Shuwei Bai, Ling Dai, Zhongxiang Ding, Jiameng Liu, Qian Wang, Qian Liu, and Dinggang Shen. “Towards General Text-guided Image Synthesis for Customized Multimodal Brain MRI Generation”. In: arXiv preprint arXiv:2409.16818 (2024) (page 3)

  38. [46]

    Image quality as- sessment: From error visibility to structural similarity

    Zhou Wang, Alan C. Bovik, Hamid R. Sheikh, and Eero P. Simoncelli. “Image quality as- sessment: From error visibility to structural similarity”. In: IEEE Transactions on Image Processing 13.4 (2004), pp. 600–612 (page 5). 12

  39. [47]

    fastMRI: An open dataset and benchmarks for accelerated MRI

    Jure Zbontar, Florian Knoll, Anuroop Sriram, Tullie Murrell, Zhengnan Huang, Matthew J Muckley, Aaron Defazio, Ruben Stern, Patricia Johnson, Mary Bruno, et al. “fastMRI: An open dataset and benchmarks for accelerated MRI”. In: arXiv preprint arXiv:1811.08839 (2018) (pages 7, 10)

  40. [48]

    The unrea- sonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. “The unrea- sonable effectiveness of deep features as a perceptual metric”. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2018, pp. 586–595 (page 5)

  41. [49]

    fastMRI+: Clinical Pathology Annotations for Knee and Brain Fully Sampled Multi-Coil MRI Data

    Ruiyang Zhao, Burhaneddin Yaman, Yuxin Zhang, Russell Stewart, Austin Dixon, Flo- rian Knoll, Zhengnan Huang, Yvonne W Lui, Michael S Hansen, and Matthew P Lungren. “fastMRI+: Clinical Pathology Annotations for Knee and Brain Fully Sampled Multi-Coil MRI Data”. In: arXiv prepr...

Pith tools

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