Pith. sign in

REVIEW 5 major objections 5 minor 39 references

AMRG: Extend Vision Language Models for Automatic Mammography Report Generation

T0 review · 5 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A LoRA-tuned MedGemma generates narrative mammography reports and beats larger general-purpose VLMs on a public benchmark.

desk verdict A useful public benchmark for mammography report generation, undermined by loose statistics and test-set hyperparameter selection that currently make the headline comparisons unproven. read the letter →

arxiv 2508.09225 v1 pith:FL5WRBE7 submitted 2025-08-12 eess.IV cs.AIcs.CV

Nak-Jun Sung , Donghyun Lee , Bo Hwa Choi , Chae Jung Park This is my paper · ORCID
classification eess.IVcs.AIcs.CV
keywords automaticmammographyreportgenerationvision-languagemodelslow-rankadaptationMedGemmaBI-RADSDMIDdatasetparameter-efficientfine-tuningradiology
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 claims that narrative mammography reports can be generated end-to-end from paired mammograms and diagnostic reports by a domain-specialized vision-language model tuned with low-rank adapters, and that this beats both larger general-purpose VLMs and modular encoder-decoder baselines under one fixed protocol. The authors build AMRG on MedGemma-4B-it, insert LoRA adapters into all linear layers, and train on the public DMID dataset with a causal language-modeling loss. On the 52-case test split they report ROUGE-L 0.5691, METEOR 0.6152, CIDEr 0.5818, and BI-RADS accuracy 0.5582, with the configuration $r=32$, $\alpha=16$ working best. The paper also positions AMRG as the first reproducible benchmark for mammography report generation on a public paired image-report dataset. If the claim holds, the contribution is a small, reproducible recipe for adapting a medical VLM to a low-resource imaging domain without full fine-tuning.

What carries the argument

The load-bearing mechanism is the AMRG pipeline: MedGemma-4B-it with Low-Rank Adaptation (LoRA) inserted into every linear projection layer, including attention query/key/value/output projections, feed-forward layers, and gated MLP projections, while all pretrained weights stay frozen. The LoRA update is $\Delta W = \alpha AB$ with rank $r$ and scaling factor $\alpha$; only the LoRA matrices, the embedding layer, and the language-model head are trained. A causal language-modeling loss conditions report generation on the image and instruction prompt, and a preprocessing chain of Otsu thresholding, left-right flip for laterality, and CLAHE contrast enhancement standardizes multiview mammograms into 512$\times$512 inputs before they reach the model.

What would settle it

Take the same 52 test reports, replace the unspecified label-extraction step with two explicit procedures, an exact keyword match for BI-RADS phrases and a separately trained BI-RADS classifier, and recompute accuracy; if the two results diverge from 0.5582 by more than a few cases, the clinical metric is not a stable property of the model.

Watch

Extended reading notes

Core claim

The central claim is that domain specialization beats scale for this task. MedGemma-4B-it, a 4B medical instruction-tuned VLM, fine-tuned with LoRA, outperforms Qwen2.5-VL-7B on ROUGE-L, METEOR, CIDEr, word-level F1, and BI-RADS accuracy, while only narrowly trailing on BLEU-1 and ROUGE-2; the modular CLIP+GPT2 and MedCLIP+GPT2 pipelines lag substantially, and Phi-3.5-Vision performs worst. The paper interprets this hierarchy as evidence that medical pretraining and instruction tuning, combined with moderate LoRA capacity, supply clinical priors that generic visual-linguistic alignment lacks. It also claims this is the first end-to-end framework for free-text mammography report generation on a public dataset, with low scores from the frozen backbone showing that the LoRA adaptation is doing the work.

Load-bearing premise

The reported clinical accuracy numbers stand or fall on the unstated rule that converts each generated free-text report into a BI-RADS label, so the 0.5582 figure could be measuring how closely the wording matches rather than true diagnostic classification.

Editorial extensions

If this is right

  • A 4B domain-specialized VLM fine-tuned with LoRA on just 407 training cases can generate narrative reports that beat 7B and 4B general-purpose VLMs and modular CLIP+GPT2 baselines under the same tuning protocol.
  • The best configuration, rank 32 with scaling factor 16 and temperature 0.1, reaches ROUGE-L 0.5691, METEOR 0.6152, CIDEr 0.5818, and BI-RADS accuracy 0.5582 on the 52-case test set.
  • Raising the LoRA rank to 64 degrades both language and clinical metrics, indicating that overfitting, not capacity, limits performance on the small DMID dataset.
  • The frozen MedGemma baseline scores far below the adapted model on every metric, so most task ability comes from the LoRA adaptation rather than from the pretrained model alone.
  • Because DMID is public and the split is fixed, these numbers form a baseline that later work on mammography report generation can compare against directly.

Reading between the lines

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

  • Because the whole comparison runs on one public dataset with one tuning protocol, a natural next test is to apply the same LoRA recipe to another small paired imaging domain, such as ultrasound or MRI, to see whether the domain-specialized backbone keeps its advantage; the paper does not run that experiment.
  • The unstated BI-RADS extraction rule means the clinical accuracy numbers function as report-similarity scores until the rule is published; a separately trained radiology text labeler would make the metric reproducible.
  • The paper's preprocessing downsamples mammograms to 512$\times$512 even though the authors list high-resolution visual cues as a core difficulty, so feeding multi-view images at native resolution is a testable extension that might change both generation quality and the BI-RADS numbers.
  • The large gap between the frozen and adapted baselines suggests the LoRA adapters encode most of the mammography-specific vocabulary, implying that the same backbone may need retraining rather than prompting when the target reporting style changes.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper introduces AMRG, a mammography report generation framework built by applying LoRA fine-tuning to the MedGemma-4B-it vision-language model and evaluating it on the public DMID dataset. The authors report a systematic LoRA hyperparameter ablation, a comparison against four other backbones (Qwen2.5-VL-7B, Phi-3.5-Vision, CLIP+GPT2, MedCLIP+GPT2) under a supposed unified tuning protocol, and both standard NLP metrics (BLEU, ROUGE, METEOR, CIDEr, word-level F1) and clinical metrics (BI-RADS accuracy, breast density accuracy). The central claim is that AMRG is the first end-to-end framework for free-text mammography report generation and outperforms general-purpose VLMs, with headline scores of ROUGE-L 0.5691, METEOR 0.6152, CIDEr 0.5818, and BI-RADS accuracy 0.5582. The manuscript also includes image preprocessing steps, a description of the LoRA adaptation, qualitative examples, and a limitations section.

Significance. If the experimental claims hold, a lightly tuned 4B medical-domain VLM would provide a useful and reusable baseline for a publicly available mammography report dataset, and the systematic comparison of medical versus general-purpose backbones under one PEFT protocol would be informative. The paper has genuine strengths: it uses a public dataset (DMID), gives a detailed preprocessing pipeline, and demonstrates that fine-tuning is necessary because the frozen MedGemma baseline scores near zero. However, the current evaluation is not statistically grounded: there are no confidence intervals or significance tests, the clinical label extraction rule is unspecified, and several reported clinical accuracies appear to use a different denominator from the stated test split. These issues are load-bearing for the headline 'outperforms' claim, so the contribution is best viewed as a promising draft rather than an established benchmark.

major comments (5)
  1. [Section 3.1, Tables 1-2, Table A.1] The test split is fixed at 52 cases in Section 3.1 and Table A.1, but many clinical accuracies in Tables 1 and 2 are exact multiples of 1/51, including 0.4510 = 23/51, 0.3529 = 18/51, 0.1176 = 6/51, and 0.4902 = 25/51. This suggests the clinical metrics were computed on a 51-case denominator rather than the described 52-case test set. The authors must report the exact denominator for every clinical metric and reconcile this discrepancy, because a one-case difference is about 2% of the test set and is material to the BI-RADS accuracy gap between systems.
  2. [Sections 4.1 and 4.2] The hyperparameters used for all backbone comparisons were selected from Table 1, but Section 4.1 never states whether Table 1 is computed on the validation or test split. If Table 1 is test-set performance, then choosing (r = 32, alpha = 16) on the basis of that table and then reporting Table 2 on the same test set means every model is evaluated with hyperparameters tuned to the evaluation set; this biases the comparison in favor of MedGemma and invalidates the 'outperforms' claim. The authors should either use a held-out validation split for hyperparameter selection and state this explicitly, or report a corrected protocol that avoids test-set contamination.
  3. [Section 4.1] The clinical metrics are not reproducible as specified. The text only states that accuracy is 'the proportion of exact matches between predicted and ground-truth labels' and does not define how a BI-RADS or breast density label is extracted from a generated free-text report. If extraction is keyword-based, the reported clinical accuracy may measure phrasing similarity rather than clinical classification. The authors should specify the extraction procedure (rule-based parser, prompted LLM, or classifier), validate that extractor, and report its agreement with the ground-truth labels.
  4. [Tables 1-2] All conclusions rest on single point estimates with no confidence intervals, significance tests, or multiple seeds. With 52 test cases, the headline ROUGE-L gap (0.5691 vs 0.5634) and the BI-RADS gap (0.5582 vs 0.4510, roughly 5-6 cases) are within plausible sampling noise. The authors should report bootstrap or exact binomial confidence intervals for each metric, and ideally multiple seeds or a paired test for the main system comparisons.
  5. [Section 3.1 and Eq. (3)] The abstract and qualitative section emphasize multi-view reasoning, but the described method does not explain how multiple mammographic views are combined. Preprocessing in Section 3.1 is defined per image, and Eqs. (3)-(4) condition generation on a single image I. The authors should specify how CC/MLO views are fed to the model (for example, concatenation, separate vision tokens, or per-view decoding) and confirm whether the quantitative evaluation actually uses multiple views per case.
minor comments (5)
  1. [Section 4.1, text after Table 1] The prose says that (r = 32, alpha = 16) achieves ROUGE-L 0.52, METEOR 0.5194, and CIDEr 0.5336, but Table 1 reports 0.5691, 0.6152, and 0.5818 for that configuration; please correct the mismatch.
  2. [Section 3.3 heading] The heading 'Casual LM Loss' should be 'Causal LM Loss.'
  3. [Section 5.2] The word 'resurts' appears in the LoRA analysis paragraph and should be corrected to 'results.'
  4. [Abstract and Section 1] The claims 'first end-to-end framework' and 'first reproducible benchmark' are stronger than the evidence and related-work discussion support; Yalunin et al. already address free-text mammography report generation, and no code or trained weights are released. Please soften or justify these priority claims.
  5. [Section 5.3] The limitations section appropriately acknowledges that surface-level NLP metrics do not fully capture clinical correctness, but the abstract and conclusion still describe the clinical metrics as evidence of strong performance; please align the claims with this acknowledged limitation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported results are empirical comparisons against held-out ground truth, not derivations from fitted inputs.

full rationale

The paper's central claims are empirical: a LoRA-tuned MedGemma achieves particular ROUGE-L, METEOR, CIDEr, and BI-RADS accuracy values on a held-out test split of DMID. These values are obtained by decoding test images and comparing generated text to radiologist-written ground-truth reports; no equation in Section 3 defines the evaluation metrics as functions of the training objective or of the LoRA hyperparameters. The LoRA update (Eqs. 1-2) and the causal language-modeling losses (Eqs. 3-10) are standard training objectives that do not contain the reported test metrics. MedGemma is cited from an external technical report by unrelated authors, and no load-bearing self-citation, uniqueness theorem, or prior-work ansatz is invoked to force the choice of backbone. The comparisons across VLM backbones are direct measurements under a stated tuning protocol rather than reductions. The lack of an explicit BI-RADS extraction rule, the small test set, and the possibility that LoRA hyperparameters were selected using the evaluation split are experimental-reporting and statistical concerns; they do not make the derivation circular because the metrics could be wrong or unrepresentative without making any step equivalent to its own input by construction.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The central empirical claims rest on the quality of DMID ground truth, the validity of automatic metrics, the representativeness of the 52-case test set, and the transferability of MedGemma's pretrained knowledge. No new scientific entities are introduced; the only added objects are fine-tuned weights, which are not independent entities.

free parameters (7)
  • LoRA rank r = 32
    Selected by grid search over {16, 32, 64}; all backbone comparisons use r=32 and the reported headline metrics depend on this choice.
  • LoRA scaling factor alpha = 16
    Selected by grid search over {8, 16}; combined with r=32 it yields the reported best results in Tables 1 and 2.
  • LoRA dropout = 0.05
    Fixed by hand in Section 3.2 for all runs.
  • Learning rate = 1e-4
    Fixed for all runs in Section 4.1; no learning-rate sweep is reported.
  • Epochs = 20
    Fixed for all runs in Section 4.1 with no early stopping or validation-based stopping described.
  • Decoding temperature = 0.1
    Used for all generations in Tables 1 and 2; affects fluency and label extraction.
  • Image preprocessing constants = 512x512 resize, CLAHE clip 2.0, tile 8x8
    Chosen in Section 3.1; changes the effective resolution and contrast available to the vision encoder.
assumptions (4)
  • domain assumption DMID ground-truth reports and BI-RADS/density labels are clinically accurate and consistent enough to serve as training targets and evaluation labels.
    The evaluation uses these labels as ground truth throughout Section 4 and Table A.1; the authors themselves note label subjectivity in Section 5.1.
  • domain assumption NLP metrics and BI-RADS/density accuracy correlate with clinically meaningful report quality.
    Sections 4 and 5 use ROUGE, METEOR, CIDEr, and label accuracy as primary evidence; Section 5.3 concedes these metrics do not fully capture clinical correctness.
  • domain assumption The 52-case test split is large enough and representative enough to support head-to-head model comparisons.
    Tables 1 and 2 report point estimates without confidence intervals, so the comparison relies on the 52-case test set described in Section 3.1.
  • domain assumption MedGemma's medical pretraining transfers to mammography after LoRA adaptation.
    The central method freezes MedGemma-4B-it and adapts only LoRA, embedding, and output head parameters, as described in Section 3.2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AMRG: Extend Vision Language Models for Automatic Mammography Report Generation." pith.science (2026). https://pith.science/paper/FL5WRBE7

@misc{pith2026250809225,
  author       = {Pith},
  title        = {Pith review of: AMRG: Extend Vision Language Models for Automatic Mammography Report Generation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FL5WRBE7}},
  note         = {Machine review of arXiv:2508.09225}
}
read the original abstract

Mammography report generation is a critical yet underexplored task in medical AI, characterized by challenges such as multiview image reasoning, high-resolution visual cues, and unstructured radiologic language. In this work, we introduce AMRG (Automatic Mammography Report Generation), the first end-to-end framework for generating narrative mammography reports using large vision-language models (VLMs). Building upon MedGemma-4B-it-a domain-specialized, instruction-tuned VLM-we employ a parameter-efficient fine-tuning (PEFT) strategy via Low-Rank Adaptation (LoRA), enabling lightweight adaptation with minimal computational overhead. We train and evaluate AMRG on DMID, a publicly available dataset of paired high-resolution mammograms and diagnostic reports. This work establishes the first reproducible benchmark for mammography report generation, addressing a longstanding gap in multimodal clinical AI. We systematically explore LoRA hyperparameter configurations and conduct comparative experiments across multiple VLM backbones, including both domain-specific and general-purpose models under a unified tuning protocol. Our framework demonstrates strong performance across both language generation and clinical metrics, achieving a ROUGE-L score of 0.5691, METEOR of 0.6152, CIDEr of 0.5818, and BI-RADS accuracy of 0.5582. Qualitative analysis further highlights improved diagnostic consistency and reduced hallucinations. AMRG offers a scalable and adaptable foundation for radiology report generation and paves the way for future research in multimodal medical AI.

Figures

Figures reproduced from arXiv: 2508.09225 by the authors.

Figure 1
Figure 1. Overview of our proposed Automatic Mammography Report Gener [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Stages of the Mammography Image Preprocessing Pipeline. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison of generated mammography reports across five VLMs: MedGemma-4B, Qwen2.5-VL, Phi-3.5-Vision, CLIP [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 26 canonical work pages

  1. [1]

    Z. He, A. N. N. Wong, J. S. Yoo, Radiology report generation using auto- matic keyword adaptation, frequency-based multi-label classification and text-to-text large language models, Computers in Biology and Medicine 196 (2025) 110625

    work page 2025

  2. [2]

    D. A. Spak, J. Plaxco, L. Santiago, M. Dryden, B. Dogan, Bi-rads ® fifth edition: A summary of changes, Diagnostic and interventional imaging 98 (3) (2017) 179–190

    work page 2017

  3. [3]

    Korean Breast Cancer Society, Breast Cancer Facts & Figures 2024, Ko- rean Breast Cancer Society, Seoul, 2024

    work page 2024

  4. [4]

    Z. Chen, Y . Song, T.-H. Chang, X. Wan, Generating radiology reports via memory-driven transformer, arXiv preprint arXiv:2010.16056 (2020)

    arXiv 2020

  5. [5]

    H. Jin, H. Che, Y . Lin, H. Chen, Promptmrg: Diagnosis-driven prompts for medical report generation, in: Proceedings of the AAAI Conference on Artificial Intelligence, V ol. 38, 2024, pp. 2607–2615

    work page 2024

  6. [6]

    A. Smit, S. Jain, P. Rajpurkar, A. Pareek, A. Y . Ng, M. P. Lungren, Chexbert: combining automatic labelers and expert annotations for accu- rate radiology report labeling using bert, arXiv preprint arXiv:2004.09167 (2020)

    arXiv 2020

  7. [7]

    Sellergren, S

    A. Sellergren, S. Kazemzadeh, T. Jaroensri, A. Kiraly, M. Traverse, T. Kohlberger, S. Xu, F. Jamil, C. Hughes, C. Lau, J. Chen, F. Mah- var, L. Yatziv, T. Chen, B. Sterling, S. A. Baby, S. M. Baby, J. Lai, S. Schmidgall, L. Yang, K. Chen, P. Bjornsson, S. Reddy, R. Brush, K. Philbrick, H. Hu, H. Yang, R. Tiwari, S. Jansen, P. Singh, Y . Liu, S. Azizi, A. ...

    arXiv 2025

  8. [8]

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, W. Chen, et al., Lora: Low-rank adaptation of large language models., ICLR 1 (2) (2022) 3

    2022

Show all 39 references
  1. [9]

    Papineni, S

    K. Papineni, S. Roukos, T. Ward, W.-J. Zhu, Bleu: a method for auto- matic evaluation of machine translation, in: Proceedings of the 40th an- nual meeting of the Association for Computational Linguistics, 2002, pp. 311–318

    2002

  2. [10]

    Lin, Rouge: A package for automatic evaluation of summaries, in: Text summarization branches out, 2004, pp

    C.-Y . Lin, Rouge: A package for automatic evaluation of summaries, in: Text summarization branches out, 2004, pp. 74–81

    2004

  3. [11]

    Banerjee, A

    S. Banerjee, A. Lavie, Meteor: An automatic metric for mt evaluation with improved correlation with human judgments, in: Proceedings of the acl workshop on intrinsic and extrinsic evaluation measures for machine translation and/or summarization, 2005, pp. 65–72

    2005

  4. [12]

    Vedantam, C

    R. Vedantam, C. Lawrence Zitnick, D. Parikh, Cider: Consensus-based image description evaluation, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 4566–4575

    2015

  5. [13]

    P. Oza, U. Oza, R. Oza, P. Sharma, S. Patel, P. Kumar, B. Gohel, Digital mammography dataset for breast cancer diagnosis research (dmid) with breast mass segmentation analysis, Biomedical Engineering Letters 14 (2) (2024) 317–330

    2024

  6. [14]

    Bannur, S

    S. Bannur, S. Hyland, Q. Liu, F. Perez-Garcia, M. Ilse, D. C. Castro, B. Boecking, H. Sharma, K. Bouzid, A. Thieme, et al., Learning to ex- ploit temporal structure for biomedical vision-language processing, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pat...

    2023

  7. [15]

    Z. Wang, Z. Wu, D. Agarwal, J. Sun, Medclip: Contrastive learning from unpaired medical images and text, in: Proceedings of the Conference on Empirical Methods in Natural Language Processing. Conference on Empirical Methods in Natural Language Processing, V ol. 2022, 2022, p. 3876

    2022

  8. [16]

    Huang, L

    S.-C. Huang, L. Shen, M. P. Lungren, S. Yeung, Gloria: A multimodal global-local representation learning framework for label-e fficient medi- cal image recognition, in: Proceedings of the IEEE /CVF international conference on computer vision, 2021, pp. 3942–3951

    2021

  9. [17]

    A. E. Johnson, T. J. Pollard, S. J. Berkowitz, N. R. Greenbaum, M. P. Lungren, C.-y. Deng, R. G. Mark, S. Horng, Mimic-cxr, a de-identified publicly available database of chest radiographs with free-text reports, Scientific data 6 (1) (2019) 317

    2019

  10. [18]

    C. Li, C. Wong, S. Zhang, N. Usuyama, H. Liu, J. Yang, T. Naumann, H. Poon, J. Gao, Llava-med: Training a large language-and-vision assis- tant for biomedicine in one day, Advances in Neural Information Process- ing Systems 36 (2023) 28541–28564

    2023

  11. [19]

    H. Liu, C. Li, Q. Wu, Y . J. Lee, Visual instruction tuning, Advances in neural information processing systems 36 (2023) 34892–34916

    2023

  12. [20]

    C. Wu, X. Zhang, Y . Zhang, Y . Wang, W. Xie, Towards generalist foun- dation model for radiology by leveraging web-scale 2d&3d medical data, arXiv preprint arXiv:2308.02463 (2023)

    2023 arXiv

  13. [21]

    M. Z. Hossain, F. Sohel, M. F. Shiratuddin, H. Laga, A comprehensive survey of deep learning for image captioning, ACM Computing Surveys (CsUR) 51 (6) (2019) 1–36

    2019

  14. [22]

    Q. You, H. Jin, Z. Wang, C. Fang, J. Luo, Image captioning with semantic attention, in: Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 4651–4659

    2016

  15. [23]

    J. Yuan, H. Liao, R. Luo, J. Luo, Automatic radiology report genera- tion based on multi-view image fusion and medical concept enrichment, in: International conference on medical image computing and computer- assisted intervention, Springer, 2019, pp. 721–729

    2019

  16. [24]

    Z. Wang, L. Liu, L. Wang, L. Zhou, R2gengpt: Radiology report genera- tion with frozen llms, Meta-Radiology 1 (3) (2023) 100033

    2023

  17. [25]

    Pellegrini, E

    C. Pellegrini, E. ¨Ozsoy, B. Busam, N. Navab, M. Keicher, Radialog: A large vision-language model for radiology report generation and conver- sational assistance, arXiv preprint arXiv:2311.18681 (2023)

    2023 arXiv

  18. [26]

    A. M. Khan, M. M. Mohsan, M. U. Akram, T. Hassan, S. G. Khawaja, A. Qayyum, Radiology report generation from a singular perspective us- ing transformers with knowledge distillation, Biomedical Signal Process- ing and Control 111 (2026) 108340

    2026

  19. [27]

    Irvin, P

    J. Irvin, P. Rajpurkar, M. Ko, Y . Yu, S. Ciurea-Ilcus, C. Chute, H. Mark- lund, B. Haghgoo, R. Ball, K. Shpanskaya, et al., Chexpert: A large chest radiograph dataset with uncertainty labels and expert comparison, in: Pro- ceedings of the AAAI conference on artificial intelli...

    2019

  20. [28]

    S ˆırbu, I.-R

    I. S ˆırbu, I.-R. Sˆırbu, J. Bogojeska, T. Rebedea, Git-cxr: End-to-end trans- former for chest x-ray report generation (2025). arXiv:2501.02598. URL https://arxiv.org/abs/2501.02598

    2025 arXiv

  21. [29]

    Singh, S

    P. Singh, S. Singh, Chestx-transcribe: a multimodal transformer for auto- mated radiology report generation from chest x-rays, Frontiers in Digital Health 7 (2025) 1535168

    2025

  22. [30]

    K. Liu, Z. Ma, X. Kang, Y . Li, K. Xie, Z. Jiao, Q. Miao, Enhanced con- trastive learning with multi-view longitudinal data for chest x-ray report generation, in: Proceedings of the Computer Vision and Pattern Recogni- 9 tion Conference, 2025, pp. 10348–10359

    2025

  23. [31]

    Zhang, B

    L. Zhang, B. Yun, Q. Li, Y . Wang, Historical report guided bi-modal concurrent learning for pathology report generation, arXiv preprint arXiv:2506.18658 (2025)

    2025 arXiv

  24. [32]

    J. Lei, X. Zhang, C. Wu, L. Dai, Y . Zhang, Y . Zhang, Y . Wang, W. Xie, Y . Li, Autorg-brain: Grounded report generation for brain mri, arXiv preprint arXiv:2407.16684 (2024)

    2024 arXiv

  25. [33]

    Sawyer-Lee, F

    R. Sawyer-Lee, F. Gimenez, A. Hoogi, D. Rubin, Curated breast imaging subset of digital database for screening mammography (cbis-ddsm), (No Title) (2016)

    2016

  26. [34]

    H. T. Nguyen, H. Q. Nguyen, H. H. Pham, K. Lam, L. T. Le, M. Dao, V . Vu, Vindr-mammo: A large-scale benchmark dataset for computer- aided diagnosis in full-field digital mammography, Scientific Data 10 (1) (2023) 277

    2023

  27. [35]

    Yalunin, E

    A. Yalunin, E. Sokolova, I. Burenko, A. Ponomarchuk, O. Puchkova, D. Umerenkov, Generating mammography reports from multi-view mam- mograms with BERT, in: M.-F. Moens, X. Huang, L. Specia, S. W.-t. Yih (Eds.), Findings of the Association for Computational Linguistics: EMNLP 20...

    2021 doi

  28. [36]

    Otsu, A threshold selection method from gray-level histograms, IEEE Transactions on Systems, Man, and Cybernetics 9 (1) (1979) 62–66

    N. Otsu, A threshold selection method from gray-level histograms, IEEE Transactions on Systems, Man, and Cybernetics 9 (1) (1979) 62–66

    1979

  29. [37]

    S. Bai, K. Chen, X. Liu, J. Wang, W. Ge, S. Song, K. Dang, P. Wang, S. Wang, J. Tang, H. Zhong, Y . Zhu, M. Yang, Z. Li, J. Wan, P. Wang, W. Ding, Z. Fu, Y . Xu, J. Ye, X. Zhang, T. Xie, Z. Cheng, H. Zhang, Z. Yang, H. Xu, J. Lin, Qwen2.5-vl technical report (2025). arXiv: 250...

    2025 arXiv

  30. [38]

    Abdin, J

    M. Abdin, J. Aneja, H. Awadalla, A. Awadallah, A. A. Awan, N. Bach, A. Bahree, A. Bakhtiari, J. Bao, H. Behl, A. Benhaim, M. Bilenko, J. Bjorck, S. Bubeck, M. Cai, Q. Cai, V . Chaudhary, D. Chen, D. Chen, W. Chen, Y .-C. Chen, Y .-L. Chen, H. Cheng, P. Chopra, X. Dai, M. Dixon...

    2024 arXiv

  31. [39]

    Radford, J

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark, et al., Learning transferable visual models from natural language supervision, in: International con- ference on machine learning, PmLR, 2021, pp. 8748–8763. Appendix ...

    2021

Pith tools

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