Pith. sign in

REVIEW 3 major objections 6 minor 39 references

MRI Image Generation Based on Text Prompts

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This paper claims that text-prompted MRI generation is feasible and practically useful: fine-tuning Stable Diffusion on paired MRI-text data yields T1, T2, and FLAIR brain images, and synthetic images can raise contrast-classification…

desk verdict The generation results are plausible, but the headline augmentation claim confounds synthetic data with doubled sample size and lacks repeated runs, so the 96.15% figure is not yet pinned down. read the letter →

arxiv 2505.22682 v1 pith:JUI5F64F submitted 2025-05-23 eess.IV cs.CVphysics.med-ph

Xinxian Fan , Mengye Lyu This is my paper · ORCID
classification eess.IVcs.CVphysics.med-ph
keywords MagneticResonanceImagesImageGenerationDiffusionModelTextGuidanceStableDataAugmentationMRIContrastClassificationLow-Field
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

MRI datasets are costly, scarce, and privacy-restricted, which starves medical AI of training data. The paper argues that text-prompted image generation can help: a Stable Diffusion model fine-tuned on paired MRI-text data produces plausible T1, T2, and FLAIR brain slices from short descriptions of field strength, contrast, and slice position. Adding thirteen synthetic slices per class to a small real training set lifts a three-class contrast classifier from 71.15% to 96.15% accuracy in the paper's test, suggesting synthetic data can substitute for scarce real acquisitions. A sympathetic reading is that the contribution is a demonstration of feasibility plus a recipe: dataset construction, two fine-tuning strategies, and two evaluation metrics, rather than a new generative architecture.

What carries the argument

The mechanism that carries the argument is the cross-attention-conditioned UNet in Stable Diffusion, which injects text-prompt embeddings into the denoising latent process, so the generator learns to associate phrases like "0.3T brain MRI, slice 5, T1 contrast" with image content. Two fine-tuning strategies operationalize this: DreamBooth (full-parameter fine-tuning with a rare identifier plus class noun, which requires a single shared prompt per batch) and direct UNet fine-tuning with per-slice structured prompts. FID, computed with both InceptionV3 and CLIP encoders, measures distributional closeness and semantic consistency of the generated set to real images; MS-SSIM measures structural similarity among generated images, hence inversely their diversity. The downstream demonstration freezes ResNet50's convolutional backbone and trains only the fully connected layer, a design choice that limits overfitting in the scarce-sample setting.

What would settle it

A reader could run the same four training-set conditions on a larger, multi-scanner test set (say 300+ slices from several 0.3–0.35T devices with mixed patient ages). If the accuracy gain from adding synthetic images does not reproduce, or if synthetic-only training falls to chance, the paper's central claim is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that the Stable Diffusion architecture, originally trained on natural images, can be repurposed into a controllable MRI generator by fine-tuning on paired image-text data from the 0.3T M4Raw and 3T fastMRI datasets. Two fine-tuning routes are compared: DreamBooth, which binds generic modality and field-strength concepts to rare identifiers, and direct fine-tuning of the UNet denoiser with structured prompts encoding field strength, slice index, and contrast type. Both routes improve on the unmodified model: IncepV3-FID falls from 317.35 to 88.18 at the best configuration, CLIP-FID from 39.94 to 9.98, and MS-SSIM rises from 0.05 to 0.29–0.38, indicating better fidelity at some cost in diversity. The downstream experiment is the strongest claim: with only one real subject's 13 slices per contrast as training data, adding 13 generated slices per contrast raises a frozen-backbone ResNet50 classifier's test accuracy on a 156-image 0.35T test set from 71.15% to 96.15%, which the authors take as evidence that generated images carry label-relevant contrast structure.

Load-bearing premise

The load-bearing assumption is that 13 generated slices from a model trained on healthy 0.3T volunteers can transfer contrast-defining structure to a test set from a different scanner, field strength, and patient population (elderly patients), and that a single train-test split with perfect validation accuracy is enough to trust the 96.15% result.

Editorial extensions

If this is right

  • Adding synthetic low-field MRI slices to a small real training set can raise contrast-classification accuracy from 71.15% to 96.15%, so text-prompted generation is a viable data-augmentation route when real MRI data are scarce.
  • Synthetic-only training nearly matches real-data training (69.87% vs 71.15% accuracy), meaning generated images retain enough contrast-specific structure to stand in for real images when none are available.
  • Fine-tuning with structured text prompts yields FID improvements across all six field-strength and contrast combinations (IncepV3-FID down to 70–133 depending on setting), indicating the recipe generalizes across acquisition parameters.
  • Because the model can be prompted with slice index, it learns slice-position-specific anatomy, enabling dataset expansion that preserves anatomical plausibility at chosen positions.

Reading between the lines

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

  • A control experiment the paper does not run would distinguish generic regularization from synthetic-data value: training with 13 additional real slices from a matched scanner instead of 13 synthetic slices, and checking whether the accuracy gain persists.
  • The paper's prompt-variation results (adding unseen features to the prompt barely changes the output) imply the model's text conditioning is shallow; unfreezing the text encoder or training with more diverse prompts would be a testable extension.
  • Because the generator trains on healthy young volunteers, augmenting datasets of elderly or diseased patients with these synthetic images risks biasing classifiers toward the healthy distribution; a clinical pipeline would need to measure that bias.
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

3 major / 6 minor

Summary. The paper fine-tunes Stable Diffusion v1.5 on brain MRI data from the 0.3T M4Raw dataset and the 3T fastMRI dataset, using text prompts that specify field strength, modality, and slice position under two fine-tuning strategies (DreamBooth and UNet fine-tuning). It evaluates generation quality with Inception-V3 FID, CLIP-based FID, and MS-SSIM, and then tests downstream utility by training a ResNet50 classifier on a small 0.35T clinical dataset, with and without synthetic 0.3T images added to the training set. The authors report that fine-tuning substantially improves FID over the original model and that augmenting the 0.35T training set with synthetic images raises contrast-classification accuracy from 71.15% to 96.15%, concluding that text-prompted MRI generation is feasible and can serve as a useful tool for medical AI applications.

Significance. If the downstream claim were established, the paper would provide practical evidence that synthetic MRI can augment scarce clinical training sets, which is a relevant and timely direction. The experiment's use of an external 0.35T dataset from a different scanner and patient population than the generative training data is a genuine strength, as is the evaluation of generation quality with two complementary FID variants. However, the classification experiment that carries the central claim confounds synthetic data with doubled sample size and lacks any statistical controls, and Section 4.3.3 shows that the text-prompt component has little observable effect. The generation-quality improvement itself is credible and could be of interest, but it is of incremental value without the downstream validation.

major comments (3)
  1. [§5.2, Table 5-1; §5.4.2, Table 5-2] The key comparison that supports the paper's central claim confounds synthetic content with doubled sample size: the condition '0.35T Real + Synthetic Data' has 26 images per class while the '0.35T Real Data' condition has 13, and the only other larger condition ('3T Real Data', 18 per class) also changes the domain. There is no control condition that adds an equivalent number of real 0.35T images to the 13-image training set, so the accuracy gain from 71.15% to 96.15% cannot be attributed specifically to the synthetic images. This is load-bearing for the conclusion that generated images can augment scarce clinical training sets.
  2. [§5.3, §5.4.1, §5.4.2] The classification experiment is a single run on a single subject-level split, with one training subject, four test subjects, no repeated seeds, no cross-validation, and no statistical significance test. The validation split is only 20% of a very small training set, and §5.4.1 itself notes that 100% validation accuracy may indicate overfitting. With 13–26 training images per class and a 156-image test set, the 25-percentage-point difference between the best and worst conditions could plausibly arise from sampling variability or initialization luck; the manuscript should report multiple splits and seeds with confidence intervals or an appropriate statistical test.
  3. [§4.3.3, Figures 4-11 to 4-13] The manuscript reports that modifying the text prompt with imaging requirements or with features not present during training has little effect on the generated images, and that the same result holds for the UNet fine-tuned models. This directly weakens the 'text-prompted' element that is central to the paper's title and abstract: the observed improvements after fine-tuning could come from learning the domain distribution with fixed class labels rather than from semantic text control. The authors should either demonstrate measurable text-prompt control (for example, through a quantitative text-image alignment metric) or reframe the claims as class-conditional rather than text-prompted generation.
minor comments (6)
  1. [Chapter 4 heading] The heading '4. Experimental Results and and Discussion' contains a duplicated 'and'; the manuscript would benefit from a thorough editorial pass, as many sections read like a raw machine translation.
  2. [§3.3, Tables 4-2 to 4-4] It is not specified how the aggregate FID and MS-SSIM values in Table 4-2 are computed (for example, pooled features over all modalities and field strengths, or averaged per-class values), which makes it difficult to compare them with the per-modality tables that follow.
  3. [§4.4] No information is given about the number of generated images used for the FID and MS-SSIM computations, or about the random seeds used for sampling; reporting these details would substantially improve reproducibility.
  4. [§5.2, Table 5-1] The synthetic data for the classification experiment are generated with random slice numbers between 1 and 18, but the M4Raw-derived training distribution contains anatomy that varies across slice positions; the paper does not discuss whether this random slice assignment could affect the label-relevant contrast information.
  5. [§3.1.2] The choice to retain only the first 10 fastMRI slices and all 18 M4Raw slices is described as filtering unusable scan layers, but no quantitative evidence is provided that this threshold captures a representative slice range; a figure or histogram of selected versus discarded slices would be helpful.
  6. [§6.2] Section 6.2 acknowledges that only the UNet was fine-tuned and the text encoder was left frozen; given the weak prompt sensitivity documented in §4.3.3, the discussion could more explicitly connect this architectural choice to the prompt-control result.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the central claims are empirical and externally validated; the sole self-citation (M4Raw dataset) is not load-bearing.

full rationale

The paper's derivation chain is empirical rather than formal. The generative results are evaluated with FID and MS-SSIM against held-out M4Raw and fastMRI test slices, not against the training objective, so the quality improvements are not self-referential by construction. The downstream classification experiment is the closest thing to a 'prediction' in the paper, and it is an external check: the 0.35T test set (Anke scanner, elderly patients, Shaanxi) is disjoint in scanner brand, field strength, and population from the synthetic 0.3T training images (Xingaoyi scanner, healthy volunteers), so the reported improvement from 71.15% to 96.15% is not mathematically forced by the generative model's training data. The only self-citation is M4Raw (Lyu et al., 2023), used as a concrete open dataset; it is real data rather than an unverified premise. The manuscript itself flags the small-validation overfitting risk (Section 5.4.1) and the domain mismatch (Section 5.4.2), and Section 4.3.3 shows weak prompt sensitivity; these are correctness or robustness limitations that could undermine the strength of the claims, but they are not circularity. The augmentation comparison does confound synthetic content with doubled sample size, yet that is an experimental-design confound, not a constructional equivalence. No equation in the paper reduces to its own input, no fitted parameter is renamed as a prediction, and no uniqueness argument is imported from the authors' prior work.

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

The central claims rest on transfer-learning assumptions and small-sample evaluation protocols, not on formal axioms. The free parameters are hyperparameters and data filters, none are fitted constants in a physical derivation, and no new physical entities are proposed.

free parameters (4)
  • Best UNet fine-tuning configuration (learning rate and step count) = 1e-4 learning rate, 1000 training steps
    The synthetic augmentation set in Section 5.2 is generated with the 1e-4/1000-step model, selected by FID rather than by downstream classification; other configurations give different FID and MS-SSIM values.
  • Slice retention threshold for fastMRI = 10 slices per volume
    Set by hand in Section 3.1.2 to remove blank or partial skull slices; this changes the training distribution and may bias generation toward central anatomy.
  • Classification hyperparameters = 5e-4 learning rate, 100 epochs, ImageNet-initialized ResNet50 with frozen convolutional layers
    Chosen by hand in Section 5.3; these choices affect the magnitude of the reported accuracy gains, and no sensitivity analysis is given.
  • DreamBooth fine-tuning schedule = 5e-6 learning rate, 400 steps, batch size 1
    Listed in Table 4-1 as hand-set hyperparameters for one of the two fine-tuning strategies; the resulting model is used as the DreamBooth baseline.
assumptions (4)
  • domain assumption Stable Diffusion v1.5 pretrained on natural images can adapt to MRI through fine-tuning.
    The entire method depends on this transfer; no formal proof is given, only empirical results.
  • domain assumption FID computed with InceptionV3 or CLIP-ViT-B-32 features and MS-SSIM are meaningful quality and diversity indicators for MRI.
    These features are trained on natural images, not MRI, so their sensitivity to field strength and contrast differences is limited.
  • ad hoc to paper Retaining only the first 10 fastMRI slices and using 18 M4Raw slices gives a representative training distribution.
    The slice filter in Section 3.1.2 is based on visual signal inspection and excludes blank slices, which changes the training data and may omit some anatomy.
  • domain assumption A classifier trained with frozen convolutional layers on 13 to 26 images per class yields stable accuracy differences.
    Small-sample training with 100% validation accuracy is treated as meaningful, but it may instead reflect overfitting or easy separability.

how reviews work

0 comments
Cite this review

Pith. "Pith review of MRI Image Generation Based on Text Prompts." pith.science (2026). https://pith.science/paper/JUI5F64F

@misc{pith2026250522682,
  author       = {Pith},
  title        = {Pith review of: MRI Image Generation Based on Text Prompts},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JUI5F64F}},
  note         = {Machine review of arXiv:2505.22682}
}
read the original abstract

This study explores the use of text-prompted MRI image generation with the Stable Diffusion (SD) model to address challenges in acquiring real MRI datasets, such as high costs, limited rare case samples, and privacy concerns. The SD model, pre-trained on natural images, was fine-tuned using the 3T fastMRI dataset and the 0.3T M4Raw dataset, with the goal of generating brain T1, T2, and FLAIR images across different magnetic field strengths. The performance of the fine-tuned model was evaluated using quantitative metrics,including Fr\'echet Inception Distance (FID) and Multi-Scale Structural Similarity (MS-SSIM), showing improvements in image quality and semantic consistency with the text prompts. To further evaluate the model's potential, a simple classification task was carried out using a small 0.35T MRI dataset, demonstrating that the synthetic images generated by the fine-tuned SD model can effectively augment training datasets and improve the performance of MRI constrast classification tasks. Overall, our findings suggest that text-prompted MRI image generation is feasible and can serve as a useful tool for medical AI applications.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 33 canonical work pages

  1. [1]

    Basic principles of magnetic resonance imaging[J]

    Van Geuns R J M, Wielopolski P A, de Bruin H G, et al. Basic principles of magnetic resonance imaging[J]. Progress in cardiovascular diseases, 1999, 42(2): 149-156

    work page 1999

  2. [2]

    Between access and privacy: challenges in sharing health data[J]

    Malin B, Goodman K. Between access and privacy: challenges in sharing health data[J]. Yearbook of medical informatics, 2018, 27(01): 055-059

    work page 2018

  3. [3]

    Medical im age data augmentation: techniques, comparisons and interpretations[J]

    Goceri E. Medical im age data augmentation: techniques, comparisons and interpretations[J]. Artificial Intelligence Review, 2023, 56(11): 12561-12605

    work page 2023

  4. [4]

    Generative adversarial networks[J]

    Goodfellow I, Pouget -Abadie J, Mirza M, et al. Generative adversarial networks[J]. Communications of the ACM, 2020, 63(11): 139-144

    work page 2020

  5. [5]

    Diffusion models in vision: A survey[J]

    Croitoru F A, Hondru V , Ionescu R T, et al. Diffusion models in vision: A survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(9): 10850-10869

    work page 2023

  6. [6]

    Conditional text image generation with diffusi on models[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Zhu Y , Li Z, Wang T, et al. Conditional text image generation with diffusi on models[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2023: 14235-14245

    work page 2023

  7. [7]

    Brain imaging generation with latent diffusion models[C]//MICCAI Workshop on Deep Generative Models

    Pinaya W H L, Tudosiu P D, Dafflon J, et al. Brain imaging generation with latent diffusion models[C]//MICCAI Workshop on Deep Generative Models. Cham: Springer Nature Switzerland, 2022: 117-126

    work page 2022

  8. [8]

    A review of medical imaging research based on diffusion models [J]

    Liu Qiagen, Guan Yu, Wu Weiwen, et al. A review of medical imaging research based on diffusion models [J]. CT Theory and Application Research, 2025, 1-19

    work page 2025

Show all 39 references
  1. [9]

    Text -to-image diffusion models in generative ai: A survey[J]

    Zhang C, Zhang C, Zhang M, et al. Text -to-image diffusion models in generative ai: A survey[J]. arXiv preprint arXiv:2303.07909, 2023

    2023 arXiv

  2. [10]

    Attngan: Fine -grained text to image generation with attentional generative adversarial networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition

    Xu T, Zhang P, Huang Q, et al. Attngan: Fine -grained text to image generation with attentional generative adversarial networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 1316-1324

    2018

  3. [11]

    Learning to generate semantic layouts for higher text-image correspondence in text -to-image synthesis[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision

    Park M, Yun J, Choi S, et al. Learning to generate semantic layouts for higher text-image correspondence in text -to-image synthesis[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. 2023: 7591-7600

    2023

  4. [12]

    Inceptionism: Going deeper into neural networks[J]

    Mordvintsev A, Olah C, Tyka M. Inceptionism: Going deeper into neural networks[J]. Google research blog, 2015, 20(14): 5

    2015

  5. [13]

    The creativity of text -to-image generation[C]//Proceedings of the 25th international academic mindtrek conference

    Oppenlaender J. The creativity of text -to-image generation[C]//Proceedings of the 25th international academic mindtrek conference. 2022: 192-202

    2022

  6. [14]

    Learning transferable visual models from natural language supervision[C]//International conference on machine learning

    Radford A, Kim J W, Hallacy C, et al. Learning transferable visual models from natural language supervision[C]//International conference on machine learning. PMLR, 2021: 8748 - 8763

    2021

  7. [15]

    Generative search engines: Initial experiments[C]//Proceedings of the International Conference on Computational Creativity

    Colton S, Smith A, Berns S, e t al. Generative search engines: Initial experiments[C]//Proceedings of the International Conference on Computational Creativity. MRI Image Generation Based on Text Prompts Page 54 2021: 237-246

    2021

  8. [16]

    Vqgan-clip: Open domain image generation and editing with natural language guidance[C]//European Conference on Computer Vision

    Crowson K, Biderman S, Kornis D, et al. Vqgan-clip: Open domain image generation and editing with natural language guidance[C]//European Conference on Computer Vision. Cham: Springer Nature Switzerland, 2022: 88-105

    2022

  9. [17]

    Diffusion models beat gans on image synthesis[J]

    Dhariwal P, Nichol A. Diffusion models beat gans on image synthesis[J]. Advances in neural information processing systems, 2021, 34: 8780-8794

    2021

  10. [18]

    High -resolution image synthesis with latent diffusion models[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Rombach R, Blattmann A, Lorenz D, et al. High -resolution image synthesis with latent diffusion models[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2022: 10684-10695

    2022

  11. [19]

    Medgen3d: A deep generative framework for paired 3d image and mask generation[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention

    Han K, Xiong Y , You C, et al. Medgen3d: A deep generative framework for paired 3d image and mask generation[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham: Springer Nature Switzerland, 2023: 759-769

    2023

  12. [20]

    Roentgen: visi on-language foundation model for chest x-ray generation[J]

    Chambon P, Bluethgen C, Delbrouck J B, et al. Roentgen: visi on-language foundation model for chest x-ray generation[J]. arXiv preprint arXiv:2211.12737, 2022

    2022 arXiv

  13. [21]

    Augmenting medical image classifiers with synthetic data from latent diffusion models[J]

    Sagers L W, Diao J A, Melas-Kyriazi L, et al. Augmenting medical image classifiers with synthetic data from latent diffusion models[J]. arXiv preprint arXiv:2308.12453, 2023

    2023 arXiv

  14. [22]

    Medisyn: Text-guided diffusion models for broad medical 2d and 3d image synthesis[J]

    Cho J, Zakka C, Kaur D, et al. Medisyn: Text-guided diffusion models for broad medical 2d and 3d image synthesis[J]. arXiv preprint arXiv:2405.09806, 2024

    2024 arXiv

  15. [23]

    Shentu J, Al Moubayed N. CXR -IRGen: An Integrated Vision and Language Model for the Generation of Clinically Accurate Chest X-Ray Image-Report Pairs[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2024: 5212-5221

    2024

  16. [24]

    A foundation model utilizing chest CT volumes and radiology reports for supervised -level zero -shot detection of abnormalities[J]

    Ethem Hamamci I, Er S, Almas F, et al. A foundation model utilizing chest CT volumes and radiology reports for supervised -level zero -shot detection of abnormalities[J]. arXiv preprint arXiv: 2403.17834, 2024

    2024

  17. [25]

    LDPM: Towards undersampled MRI reconstruction with MR- V AE and Latent Diffusion Prior[J]

    Tang X, Guan J, Li L, et al. LDPM: Towards undersampled MRI reconstruction with MR- V AE and Latent Diffusion Prior[J]. arXiv preprint arXiv:2411.02951v2, 2025

    2025 arXiv

  18. [26]

    Bayesian MRI reconstruction with joint uncertainty estimation using diffusion models[J]

    Luo G, Blumenthal M, Heide M, et al. Bayesian MRI reconstruction with joint uncertainty estimation using diffusion models[J]. Magnetic Resonance in Medicine, 2023, 90(1): 295-311

    2023

  19. [27]

    Tumor detection by nuclear magnetic resonance[J]

    Damadian R. Tumor detection by nuclear magnetic resonance[J]. Science, 1971, 171(3976): 1151-1153

    1971

  20. [28]

    Image formation by induced local interactions: examples employing nuclear magnetic resonance[J]

    Lauterbur P C. Image formation by induced local interactions: examples employing nuclear magnetic resonance[J]. nature, 1973, 242(5394): 190-191

    1973

  21. [29]

    NMR 'diffraction' in solids?[J]

    Mansfield P, Grannell P K. NMR 'diffraction' in solids?[J]. Journal of Physics C: solid state physics, 1973, 6(22): L422

    1973

  22. [30]

    Principles of Magnetic Resonance Imaging [M]

    Xiong Guoxin, Li Liben. Principles of Magnetic Resonance Imaging [M]. Beijing: Science Press, 2007

    2007

  23. [31]

    Denoising diffusion probabilistic models[J]

    Ho J, Jain A, Abbeel P. Denoising diffusion probabilistic models[J]. Advances in neural information processing systems, 2020, 33: 6840-6851. MRI Image Generation Based on Text Prompts Page 55

    2020

  24. [32]

    Denoising diffusion implicit models[J]

    Song J, Meng C, Ermon S. Denoising diffusion implicit models[J]. arXiv preprint arXiv:2010.02502, 2020

    2010 arXiv

  25. [33]

    M4Raw: A multi-contrast, multi-repetition, multi-channel MRI k-space dataset for low-field MRI research[J]

    Lyu M, Mei L, Huang S, et al. M4Raw: A multi-contrast, multi-repetition, multi-channel MRI k-space dataset for low-field MRI research[J]. Scientific Data, 2023, 10(1): 264

    2023

  26. [34]

    fastMRI: An open dataset and benchmarks for accelerated MRI[J]

    Zbontar J, Knoll F, Sriram A, et al. fastMRI: An open dataset and benchmarks for accelerated MRI[J]. arXiv preprint arXiv:1811.08839, 2018

    2018 arXiv

  27. [35]

    Dreambooth: Fine tuning text-to-image diffusion models for subject -driven generation[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Ruiz N, Li Y , Jampani V , et al. Dreambooth: Fine tuning text-to-image diffusion models for subject -driven generation[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2023: 22500-22510

    2023

  28. [36]

    GANs trained by a two time -scale update rule converge to a local nash equilibrium[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems

    Heusel M, Ramsauer H, Unterthiner T, et al. GANs trained by a two time -scale update rule converge to a local nash equilibrium[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. 2017: 6629-6640

    2017

  29. [37]

    Multiscale structural similarity for image quality assessment[C]//The Thirty-Seventh Asilomar Conference on Signals, Systems & Computers, 2003: 1398-1402

    Wang Z, Simoncelli E P, Bovik A C. Multiscale structural similarity for image quality assessment[C]//The Thirty-Seventh Asilomar Conference on Signals, Systems & Computers, 2003: 1398-1402

    2003

  30. [38]

    RadBERT: adapting transformer-based language models to radiology[J]

    Yan A, McAuley J, Lu X, et al. RadBERT: adapting transformer-based language models to radiology[J]. Radiology: Artificial Intelligence, 2022, 4(4): e210258

    2022

  31. [39]

    Self-improving generative foundation model for synthetic medical image generation and clinical applications[J]

    Wang J, Wang K, Yu Y , et al. Self-improving generative foundation model for synthetic medical image generation and clinical applications[J]. Nature Medicine, 2025, 31(2): 609-617

    2025

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.