Pith. sign in

REVIEW 3 major objections 6 minor 60 references

Grounding DINO-US-SAM: Text-Prompted Multi-Organ Segmentation in Ultrasound with LoRA-Tuned Vision-Language Models

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Text prompts alone can localize and segment multiple ultrasound organs: with LoRA-tuned Grounding DINO supplying boxes to a frozen SAM2, the pipeline outperforms specialized medical segmentation methods on most seen datasets and holds up…

desk verdict Useful engineering with a real but incremental novelty and an above-average evaluation; needs table fixes and one training-description contradiction resolved before the numbers can be trusted. read the letter →

arxiv 2506.23903 v3 pith:OUMQU32W submitted 2025-06-30 cs.CV cs.AI

Hamza Rasaee , Taha Koleilat , Hassan Rivaz This is my paper · ORCID
classification cs.CVcs.AI
keywords UltrasoundimagesegmentationPrompt-drivenVision-languagemodelsGroundingDINOSegmentAnythingModel2Low-RankAdaptationMulti-organDomaingeneralization
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 argues that a free-form text prompt—just a phrase naming an organ or lesion—is enough to drive automatic segmentation in ultrasound, with no clicks, bounding boxes, or per-organ retraining. The method pairs Grounding DINO, a text-grounded object detector, with SAM2, a general-purpose image segmenter, and adapts only the detector to ultrasound using Low-Rank Adaptation, or LoRA, which trains low-rank matrices in place of full weight updates. Across 15 public ultrasound datasets used for tuning and 3 held out for testing, the authors report that the pipeline outperforms UniverSeg, MedSAM, MedCLIP-SAM, MedCLIP-SAMv2, BiomedParse, and SAMUS on most seen test sets, while retaining strong accuracy on unseen breast, thyroid, and paraspinal-muscle data. Only about 1.7% of Grounding DINO's parameters are updated and SAM2 stays entirely frozen, so the adaptation is cheap. If the claim holds, ultrasound analysis could be driven by the words of the operator rather than by organ-specific models or manual prompts.

What carries the argument

The machinery is LoRA (Low-Rank Adaptation), which represents each weight update as the product of two small matrices, $\Delta W = BA$, inserted into Grounding DINO's feature enhancer, cross-modality decoder, and frozen BERT text encoder so that only about 1.7% of detector parameters train. A composite loss—L1 and GIoU for box regression plus a focal contrastive loss for aligning text queries with visual features—steers the adaptation. The load-bearing assertion is that these LoRA adapters realign natural-image text embeddings to ultrasound appearance well enough that a completely frozen SAM2 can produce accurate masks from the proposed boxes.

What would settle it

Hold out an organ class absent from all 18 datasets—for example, fetal anatomy, the brachial plexus, or cardiac chambers—prompt with that organ name alone, and compare the resulting masks with expert annotations. If the Dice score falls far below the seen-organ range, or if near-synonym prompts produce large swings in accuracy, the claimed open-vocabulary generalization is bounded by the training distribution.

Watch

Extended reading notes

Core claim

The central claim is a division of labour between a language-aware detector and a frozen segmenter: after LoRA tuning, Grounding DINO turns a prompt such as 'tumor' or 'malignant lesion' into a tight bounding box on an ultrasound frame, and SAM2—with no ultrasound-specific training at all—converts that box into a pixel-level mask. The ablation isolates what is doing the work: unfine-tuned Grounding DINO with SAM2 performs poorly; fine-tuning Grounding DINO and pairing it with the medical-image MedSAM improves results; and fine-tuning Grounding DINO while leaving SAM2 frozen gives the best Dice and IoU scores on breast, liver, and prostate test sets. The authors read this as evidence that the ultrasound-specific bottleneck is language-to-image grounding, not mask decoding, and that SAM2 already has enough texture and anatomical context to segment accurately once it receives a correct box. The held-out results extend the claim: on breast, thyroid, and paraspinal-muscle datasets never seen in training, the same prompt-driven pipeline reaches Dice scores near or above the specialized baselines.

Load-bearing premise

The pipeline stands on the assumption that a frozen SAM2, trained on natural images and video, can segment ultrasound structures accurately once Grounding DINO's LoRA-tuned boxes are right, and that this text-to-box alignment transfers to organs and wording never seen in training.

Editorial extensions

If this is right

  • One trained model could serve as a general ultrasound segmentation front end, since a single LoRA-tuning run covers six organ systems and several lesion types.
  • Fully text-driven operation eliminates the manual point or box prompts required by tools such as SAMUS and MedSAM, making the workflow automated end to end.
  • The reported 0.33 s per 800x800 image on an older Titan V suggests real-time clinical use is within reach without expensive hardware.
  • The absence of any musculoskeletal data in fine-tuning, followed by strong results on the Luminous muscle dataset, implies that unseen anatomies can be segmented by prompt alone.

Reading between the lines

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

  • If the division of labour is the real source of the gain, then swapping frozen SAM2 for another frozen segmenter while keeping the LoRA-tuned detector fixed should leave performance roughly unchanged; this would isolate whether the improvements come from the detector adaptation or the segmenter's own capacities.
  • The prompt-sensitivity results imply a practical safeguard: a lightweight prompt validator or re-ranker could catch phrasings like 'capsular layer of the kidney' before they reach the segmenter, improving worst-case accuracy.
  • A natural extension is vascular or cardiac ultrasound, where targets are small, moving, and textually less standardized; if the frozen SAM2 can handle those boxes, the open-vocabulary claim extends well beyond the six studied organ systems.
  • Because the three unseen datasets were chosen to vary in similarity to training data, one could quantify domain shift with image statistics and test whether the accuracy drop is monotone in that distance, yielding a predictive trust model for new clinical sites.
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 proposes Grounding DINO-US-SAM, a text-prompted pipeline for multi-organ ultrasound segmentation. Grounding DINO is fine-tuned with LoRA on 15 ultrasound datasets while SAM2 remains frozen, and the predicted boxes are used as prompts for SAM2. The method is evaluated on held-out test splits of the seen datasets and on three entirely unseen datasets (BUSBRA, TNSCUI, Luminous), with reported DSC/IoU comparisons against UniverSeg, BiomedParse, SAMUS, MedCLIP-SAM, and MedCLIP-SAMv2, plus ablations, prompt-sensitivity experiments, and runtime measurements.

Significance. If the reported results are correct, the paper makes a useful empirical contribution by demonstrating that a LoRA-tuned open-vocabulary detector can serve as a text-prompt box generator for a frozen SAM2 across multiple ultrasound organs, including a previously unseen musculoskeletal domain. The experimental design is in several respects sound: held-out test splits, three completely unseen datasets, paired t-tests, and a conservative comparison in which SAMUS receives ground-truth point prompts. The paper also honestly notes the SAMUS point-prompt advantage. However, the verification of the text-localization mechanism is incomplete, and several table entries are internally inconsistent, so the central claims are not yet fully supported.

major comments (3)
  1. [Section II-B and Section II-E] Section II-B states that the bounding box regression head is 'left fully trainable without LoRA,' while Section II-E states that 'Only LoRA parameters were updated, whereas all other weights remained frozen.' These statements are mutually contradictory. The 'only 1.7% via LoRA' claim and the associated parameter-efficiency argument are load-bearing; if the box head is indeed trained, the total number of updated parameters is larger than 1.7% and the mechanism may be ordinary supervised detection rather than LoRA-based cross-modal alignment. Please clarify which parameters are updated, report the exact count of trainable parameters, and reconcile the two sections.
  2. [Table V and Section III-C] Table V reports for the BUSBRA row DSC = 79.10% and IoU = 86.44%. Since IoU ≤ DSC for any two sets, this value pair is impossible. The same numbers appear in the TNSCUI row in swapped order (DSC = 86.44%, IoU = 79.10%), and the text in Section III-C repeats the impossible pairing. These are the headline unseen-domain results, so the corrected values and a re-check of all table entries are required before the generalization claims can be assessed.
  3. [Section III-C, Table V] The paper's central claim is that text prompts alone localize organs through the LoRA-tuned Grounding DINO, but no bounding-box-level evaluation is reported anywhere. End-to-end DSC/IoU on unseen datasets cannot distinguish genuine text-prompt localization from a degenerate or full-frame box that frozen SAM2 converts into a mask, especially since the baseline Grounding DINO + SAM2 (no fine-tuning) is weak. Please report box-level metrics (e.g., box IoU or detection AP) as well as a control experiment with a full-frame box input to SAM2, or otherwise provide direct evidence that the predicted boxes are text-grounded.
minor comments (6)
  1. [Table VI] In Table VI, the 'segment kidney medulla' row reports identical DSC and IoU values (68.79 ± 18). This is implausible unless the predictions are perfect, and it is likely a typographical error; please correct it and check adjacent rows.
  2. [Table II] Table II lists the three unseen datasets (BUSBRA, TNSCUI, Luminous) in the Test column, but the Total row reports Test = 1978, which is the sum of only the seen datasets' test splits. Please clarify the total or separate seen and unseen test counts.
  3. [Section II-B text encoder] The text says the BERT-based text encoder 'remains frozen' while LoRA modules are applied to its self-attention output projections and feed-forward layers. Since LoRA parameters are trainable, the encoder is not fully frozen; please rephrase to avoid ambiguity.
  4. [Abstract and Conclusion] The abstract and conclusion claim 'real time' at 0.33 s per image; at roughly 3 frames per second this may not meet the usual clinical real-time threshold. Consider rephrasing to 'interactive-rate' or reporting the hardware-dependent caveat.
  5. [Table IV] Table IV reports a DSC of 1.57% for BiomedParse on TG3K, which is implausibly low even for a zero-shot baseline on thyroid data. Please verify that this entry and other extreme values are not transcription errors.
  6. [Section II-E] The training description mentions 'a contrastive alignment loss to optimize image-text correspondence,' but the loss function in Section II-C contains only L1, GIoU, and focal terms. Please specify where the contrastive loss is included in the final objective.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the empirical claims are evaluated on held-out and fully unseen datasets against external baselines, and self-citations are not load-bearing.

full rationale

The paper's derivation chain is empirically self-contained rather than circular. The central prediction — that LoRA-tuned Grounding DINO boxes plus frozen SAM2 achieve high DSC — is tested on per-dataset test splits (Section II-A/E, Tables IV-V) and on three datasets (BUSBRA, TNSCUI, Luminous) that are never used for training, validation, or prompt selection (Section II-A). The LoRA adapters are trained only on the seen 15 datasets with ground-truth boxes derived from masks; no parameter is fitted to the unseen test results. Ablation Table III compares frozen Grounding DINO+SAM2 against the fine-tuned version, so the conclusion that SAM2 is capable given an accurate box follows from the controlled comparison, not from a definition. Self-citations to MedCLIP-SAM/MedCLIP-SAMv2 and the Luminous database are used as external baselines or public data, not as justification for the proposed method's correctness. The apparent inconsistency between Section II-B ('bounding box regression head is left fully trainable without LoRA') and Section II-E ('Only LoRA parameters were updated') is a reporting/correctness concern about the parameter count and about which weights moved, not an instance of a prediction reducing to its input by construction. Similarly, the absence of reported intermediate boxes means the mechanism is inferred from final DSC, but inference-from-final-metrics is not circularity. Therefore no step in the claimed derivation is equivalent to its own input.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim rests on the standard foundation-model assumption that a frozen SAM2 and a LoRA-adapted Grounding DINO transfer to ultrasound, plus a set of hand-chosen hyperparameters. There are no invented entities. The main unstated degrees of freedom are the LoRA rank and prompt templates.

free parameters (6)
  • LoRA rank r
    Not reported in the paper; determines capacity of adapters and is a hand-chosen hyperparameter essential to the fine-tuning result.
  • Loss weights lambda_L1, lambda_GIoU, lambda_focal = 0.5, 0.1, 0.1
    Chosen by hand in Section II-C and used for all experiments; central to bounding box regression training.
  • Focal loss alpha and gamma = 0.25, 2.0
    Standard values taken from RetinaNet, but still hand-chosen hyperparameters in Eq. (6).
  • Learning rate and weight decay = 5e-4, 1e-4
    AdamW settings in Section II-E; selected by the authors without a search protocol.
  • Early stopping patience = 20 epochs
    Used to stop training based on validation loss; a modeling choice affecting final weights.
  • Prompt templates
    Primary prompts were designed with a radiologist; this hand-crafted input specification is part of the method and influences results, as shown in Table VI.
assumptions (5)
  • domain assumption A single pooled dataset of 15 ultrasound datasets is representative enough to teach organ-level grounding that transfers across acquisition protocols.
    The training protocol in Section II-A pools all seen datasets; if this assumption fails, generalization to unseen distributions in Section III-C would not be expected.
  • domain assumption Tight bounding boxes derived from ground-truth masks are adequate supervision for learning text-to-box grounding.
    Section II-A converts masks to minimum enclosing rectangles; detection quality depends on this proxy being faithful.
  • domain assumption Frozen SAM2, trained on natural images, can generate accurate ultrasound masks from correct boxes.
    The method relies on SAM2 with no ultrasound fine-tuning; the authors justify this only through the ablation in Table III, not through a theoretical guarantee.
  • domain assumption The text encoder of Grounding DINO, after LoRA adaptation, retains open-vocabulary competence for anatomy words not seen in ultrasound training.
    Unseen-set prompts such as 'multifidus muscle' must be parsed and grounded by the adapted text encoder; Section III-C is the only evidence.
  • ad hoc to paper LoRA rank and adapter placement described in Section II-B are sufficient to adapt cross-modal attention without destroying pretrained weights.
    The method is built on this specific adaptation choice; no ablation over rank or placement is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Grounding DINO-US-SAM: Text-Prompted Multi-Organ Segmentation in Ultrasound with LoRA-Tuned Vision-Language Models." pith.science (2026). https://pith.science/paper/OUMQU32W

@misc{pith2026250623903,
  author       = {Pith},
  title        = {Pith review of: Grounding DINO-US-SAM: Text-Prompted Multi-Organ Segmentation in Ultrasound with LoRA-Tuned Vision-Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/OUMQU32W}},
  note         = {Machine review of arXiv:2506.23903}
}
read the original abstract

Accurate and generalizable object segmentation in ultrasound imaging remains a significant challenge due to anatomical variability, diverse imaging protocols, and limited annotated data. In this study, we propose a prompt-driven vision-language model (VLM) that integrates Grounding DINO with SAM2 (Segment Anything Model2) to enable object segmentation across multiple ultrasound organs. A total of 18 public ultrasound datasets, encompassing the breast, thyroid, liver, prostate, kidney, and paraspinal muscle, were utilized. These datasets were divided into 15 for fine-tuning and validation of Grounding DINO using Low Rank Adaptation (LoRA) to the ultrasound domain, and 3 were held out entirely for testing to evaluate performance in unseen distributions. Comprehensive experiments demonstrate that our approach outperforms state-of-the-art segmentation methods, including UniverSeg, MedSAM, MedCLIP-SAM, BiomedParse, and SAMUS on most seen datasets while maintaining strong performance on unseen datasets without additional fine-tuning. These results underscore the promise of VLMs in scalable and robust ultrasound image analysis, reducing dependence on large, organ-specific annotated datasets. We will publish our code on code.sonography.ai after acceptance.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

60 extracted references · 42 canonical work pages

  1. [1]

    Fine-tuning u-net for ultrasound image segmentation: different layers, different outcomes,

    M. Amiri, R. Brooks, and H. Rivaz, “Fine-tuning u-net for ultrasound image segmentation: different layers, different outcomes,”IEEE Trans- actions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 67, no. 12, pp. 2510–2518, 2020

    work page 2020

  2. [2]

    Hctnet: A hybrid cnn-transformer network for breast ultrasound image segmentation,

    Q. He, Q. Yang, and M. Xie, “Hctnet: A hybrid cnn-transformer network for breast ultrasound image segmentation,”Computers in Biology and Medicine, vol. 155, p. 106629, 2023. AUTHORet al.: PREP ARA TION OF P APERS FOR IEEE TRANSACTIONS ON UL TRASONICS, FERROELECTRICS, AND FREQUENCY CONTROL 11 Fig. 3: Qualitative comparison of segmentation results on variou...

    work page 2023

  3. [3]

    Mcv-unet: A modified convolution & transformer hybrid encoder-decoder network with multi-scale information fusion for ultrasound image semantic segmentation,

    Z. Xu and Z. Wang, “Mcv-unet: A modified convolution & transformer hybrid encoder-decoder network with multi-scale information fusion for ultrasound image semantic segmentation,”PeerJ Computer Science, vol. 10, p. e2146, 2024

    work page 2024

  4. [4]

    Intrapartum ultrasound image segmentation of pubic sym- physis and fetal head using dual student-teacher framework with cnn-vit collaborative learning,

    J. Jiang, H. Wang, J. Bai, S. Long, S. Chen, V . M. Campello, and K. Lekadir, “Intrapartum ultrasound image segmentation of pubic sym- physis and fetal head using dual student-teacher framework with cnn-vit collaborative learning,” inInternational conference on medical image computing and computer-assisted intervention. Springer, 2024, pp. 448–458

    work page 2024

  5. [5]

    Hybrid-structure-oriented transformer for arm musculoskeletal ultra- sound segmentation,

    L. Chen, Y . Wang, Z. Zhao, H. Liao, D. Zhang, H. Han, and F. Chen, “Hybrid-structure-oriented transformer for arm musculoskeletal ultra- sound segmentation,” inInternational Conference on Medical Image Computing and Computer-Assisted Intervention. Springer, 2024, pp. 621–631

    work page 2024

  6. [6]

    Microsegnet: A deep learning approach for prostate segmentation on micro-ultrasound images,

    H. Jiang, M. Imran, P. Muralidharan, A. Patel, J. Pensa, M. Liang, T. Benidir, J. R. Grajo, J. P. Joseph, R. Terryet al., “Microsegnet: A deep learning approach for prostate segmentation on micro-ultrasound images,”Computerized Medical Imaging and Graphics, vol. 112, p. 102326, 2024

    2024

  7. [7]

    Lightbtseg: A lightweight breast tumor segmentation model using ultrasound images via dual-path joint knowledge distillation,

    H. Guo, S. Wang, H. Dang, K. Xiao, Y . Yang, W. Liu, T. Liu, and Y . Wan, “Lightbtseg: A lightweight breast tumor segmentation model using ultrasound images via dual-path joint knowledge distillation,” in 2023 China Automation Congress (CAC). IEEE, 2023, pp. 3841–3847

    work page 2023

  8. [8]

    Aau-net: an adaptive attention u-net for breast lesions segmentation in ultrasound images,

    G. Chen, L. Li, Y . Dai, J. Zhang, and M. H. Yap, “Aau-net: an adaptive attention u-net for breast lesions segmentation in ultrasound images,” IEEE Transactions on Medical Imaging, vol. 42, no. 5, pp. 1289–1300, 2022

    work page 2022

Show all 60 references
  1. [9]

    Universeg: Universal medical image segmentation,

    V . I. Butoi, J. J. G. Ortiz, T. Ma, M. R. Sabuncu, J. Guttag, and A. V . Dalca, “Universeg: Universal medical image segmentation,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 21 438–21 451

    2023

  2. [10]

    Striving for simplicity: Simple yet effective prior-aware pseudo-labeling for semi-supervised ultrasound image segmentation,

    Y . Chen, Y . Wang, Z. Zheng, J. Hu, Y . Shi, S. Xiong, X. X. Zhu, and L. Mou, “Striving for simplicity: Simple yet effective prior-aware pseudo-labeling for semi-supervised ultrasound image segmentation,” in International Conference on Medical Image Computing and Computer- As...

    2024

  3. [11]

    Ul- trasound segmentation using semi-supervised learning: Application in point-of-care sarcopenia assessment,

    H. Rasaee, M. Samuel, B. Behboodi, J. Afilalo, and H. Rivaz, “Ul- trasound segmentation using semi-supervised learning: Application in point-of-care sarcopenia assessment,”IEEE Open Journal of Engineer- ing in Medicine and Biology, 2025

    2025

  4. [12]

    Deep spectral methods for unsupervised ultrasound image interpretation,

    O. Tmenova, Y . Velikova, M. Saleh, and N. Navab, “Deep spectral methods for unsupervised ultrasound image interpretation,” inInterna- tional Conference on Medical Image Computing and Computer-Assisted Intervention. Springer, 2024, pp. 200–210

    2024

  5. [13]

    Shan: Shape guided network for thyroid nodule ultrasound cross-domain segmenta- tion,

    R. Zhang, W. Lu, C. Guan, J. Gao, X. Wei, and X. Li, “Shan: Shape guided network for thyroid nodule ultrasound cross-domain segmenta- tion,” inInternational Conference on Medical Image Computing and Computer-Assisted Intervention. Springer, 2024, pp. 732–741

    2024

  6. [14]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clarket al., “Learning transferable visual models from natural language supervision,” inInternational conference on machine learning. PmLR, 2021, pp. 8748–8763

    2021

  7. [15]

    Medclip-sam: Bridging text and image towards universal medical image segmenta- tion,

    T. Koleilat, H. Asgariandehkordi, H. Rivaz, and Y . Xiao, “Medclip-sam: Bridging text and image towards universal medical image segmenta- tion,” inInternational Conference on Medical Image Computing and Computer-Assisted Intervention. Springer, 2024, pp. 643–653

    2024

  8. [16]

    Grounding dino: Marrying dino with grounded pre-training for open-set object detection,

    S. Liu, Z. Zeng, T. Ren, F. Li, H. Zhang, J. Yang, Q. Jiang, C. Li, J. Yang, H. Suet al., “Grounding dino: Marrying dino with grounded pre-training for open-set object detection,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 38–55

    2024

  9. [17]

    Medclip: Contrastive learning from unpaired medical images and text,

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

    2022

  10. [18]

    Sam 2: Segment anything in images and videos,

    N. Ravi, V . Gabeur, Y .-T. Hu, R. Hu, C. Ryali, T. Ma, H. Khedr, R. R¨adle, C. Rolland, L. Gustafsonet al., “Sam 2: Segment anything in images and videos,”arXiv preprint arXiv:2408.00714, 2024

    2024 arXiv

  11. [19]

    A foundation model for joint segmentation, detection and recognition of biomedical objects across nine modalities,

    T. Zhao, Y . Gu, J. Yang, N. Usuyama, H. H. Lee, S. Kiblawi, T. Nau- mann, J. Gao, A. Crabtree, J. Abelet al., “A foundation model for joint segmentation, detection and recognition of biomedical objects across nine modalities,”Nature methods, vol. 22, no. 1, pp. 166–176, 2025

    2025

  12. [20]

    Medclip- samv2: Towards universal text-driven medical image segmentation,

    T. Koleilat, H. Asgariandehkordi, H. Rivaz, and Y . Xiao, “Medclip- samv2: Towards universal text-driven medical image segmentation,” arXiv preprint arXiv:2409.19483, 2024

    2024 arXiv

  13. [21]

    Segment anything in medical images,

    J. Ma, Y . He, F. Li, L. Han, C. You, and B. Wang, “Segment anything in medical images,”Nature Communications, vol. 15, no. 1, p. 654, 2024

    2024

  14. [22]

    Segment anything small for ultrasound: enhanced segmen- tation of small anatomical structures using iterative point prompts and image transformations,

    D. L. Ferreira, A. Gangopadhyay, H.-M. Chang, R. Soni, and G. Avinash, “Segment anything small for ultrasound: enhanced segmen- tation of small anatomical structures using iterative point prompts and image transformations,” inMedical Imaging 2025: Image Processing, vol. 13406....

    2025

  15. [23]

    Meyer, A

    A. Meyer, A. Murali, D. Mutter, and N. Padoy, “Ultrasam: A foundation 12 IEEE TRANSACTIONS ON UL TRASONICS, FERROELECTRICS, AND FREQUENCY CONTROL, VOL. XX, NO. XX, XXXX 2017 model for ultrasound using large open-access segmentation datasets,” arXiv preprint arXiv:2411.16222, 2024

    2017

  16. [24]

    Sam-medus: a foundational model for universal ultrasound image segmentation,

    F. Tian, J. Zhai, J. Gong, W. Lei, S. Chang, F. Ju, S. Qian, and X. Zou, “Sam-medus: a foundational model for universal ultrasound image segmentation,”Journal of Medical Imaging, vol. 12, no. 2, pp. 027 001–027 001, 2025

    2025

  17. [25]

    Clicksam: Fine-tuning segment anything model using click prompts for ultrasound image segmentation,

    A. Guo, G. Fei, H. Pasupuleti, and J. Wang, “Clicksam: Fine-tuning segment anything model using click prompts for ultrasound image segmentation,” inMedical Imaging 2024: Ultrasonic Imaging and Tomography, vol. 12932. SPIE, 2024, pp. 240–244

    2024

  18. [26]

    Beyond adapting sam: Towards end-to-end ultrasound image segmentation via auto prompting,

    X. Lin, Y . Xiang, L. Yu, and Z. Yan, “Beyond adapting sam: Towards end-to-end ultrasound image segmentation via auto prompting,” in International Conference on Medical Image Computing and Computer- Assisted Intervention. Springer, 2024, pp. 24–34

    2024

  19. [27]

    Apg-sam: Auto- matic prompt generation for sam-based breast lesion segmentation with boundary-aware optimization,

    D. Yin, Q. Zheng, L. Chen, Y . Hu, and Q. Wang, “Apg-sam: Auto- matic prompt generation for sam-based breast lesion segmentation with boundary-aware optimization,”Expert Systems with Applications, vol. 276, p. 127048, 2025

    2025

  20. [28]

    Cc-sam: Sam with cross-feature attention and context for ultrasound image segmentation,

    S. N. Gowda and D. A. Clifton, “Cc-sam: Sam with cross-feature attention and context for ultrasound image segmentation,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 108–124

    2024

  21. [29]

    Multi-organ foundation model for universal ultrasound image segmentation with task prompt and anatomical prior,

    H. Chen, Y . Cai, C. Wang, L. Chen, B. Zhang, H. Han, Y . Guo, H. Ding, and Q. Zhang, “Multi-organ foundation model for universal ultrasound image segmentation with task prompt and anatomical prior,” IEEE Transactions on Medical Imaging, 2025

    2025

  22. [30]

    Prompting foundational models for omni-supervised instance segmentation,

    A. M. Das, R. Chaudhry, K. Kundu, and D. Modolo, “Prompting foundational models for omni-supervised instance segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 1583–1592

    2024

  23. [31]

    Interpreting object-level foundation models via visual precision search,

    R. Chen, S. Liang, J. Li, S. Liu, M. Li, Z. Huang, H. Zhang, and X. Cao, “Interpreting object-level foundation models via visual precision search,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 30 042–30 052

    2025

  24. [32]

    Efficientvit-sam: Accelerated segment anything model without performance loss,

    Z. Zhang, H. Cai, and S. Han, “Efficientvit-sam: Accelerated segment anything model without performance loss,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 7859–7863

    2024

  25. [33]

    Curated benchmark dataset for ultrasound based breast lesion analysis,

    A. Pawłowska, A. ´Cwierz-Pie´nkowska, A. Domalik, D. Jagu ´s, P. Kasprzak, R. Matkowski, Ł. Fura, A. Nowicki, and N. ˙Zołek, “Curated benchmark dataset for ultrasound based breast lesion analysis,”Scientific Data, vol. 11, no. 1, p. 148, 2024

    2024

  26. [34]

    An open-access breast lesion ultrasound image database: Applicable in artificial intelligence studies,

    A. A. Ardakani, A. Mohammadi, M. Mirza-Aghazadeh-Attari, and U. R. Acharya, “An open-access breast lesion ultrasound image database: Applicable in artificial intelligence studies,”Computers in Biology and Medicine, vol. 152, p. 106438, 2023

    2023

  27. [35]

    Artificial intelligence, bi-rads evalua- tion and morphometry: A novel combination to diagnose breast cancer using ultrasonography, results from multi-center cohorts,

    H. Hamyoon, W. Y . Chan, A. Mohammadi, T. Y . Kuzan, M. Mirza- Aghazadeh-Attari, W. L. Leong, K. M. Altintoprak, A. Vijayananthan, K. Rahmat, N. Ab Muminet al., “Artificial intelligence, bi-rads evalua- tion and morphometry: A novel combination to diagnose breast cancer using ...

    2022

  28. [36]

    Applications of machine-learning algorithms for prediction of benign and malignant breast lesions using ultrasound radiomics signatures: A multi-center study,

    H. Homayoun, W. Y . Chan, T. Y . Kuzan, W. L. Leong, K. M. Altin- toprak, A. Mohammadi, A. Vijayananthan, K. Rahmat, S. S. Leong, M. Mirza-Aghazadeh-Attariet al., “Applications of machine-learning algorithms for prediction of benign and malignant breast lesions using ultrasoun...

    2022

  29. [37]

    Iqbal, “BUS UC,” Mendeley Data, V1, 2023, accessed: 2024-06-07

    A. Iqbal, “BUS UC,” Mendeley Data, V1, 2023, accessed: 2024-06-07. [Online]. Available: https://doi.org/10.17632/3ksd7w7jkx.1

    2023 doi

  30. [38]

    Bus-uclm: Breast ultrasound lesion segmentation dataset,

    N. Vallez, G. Bueno, O. Deniz, M. A. Rienda, and C. Pastor, “Bus-uclm: Breast ultrasound lesion segmentation dataset,”Scientific Data, vol. 12, no. 1, p. 242, 2025

    2025

  31. [39]

    Automated breast ultrasound lesions detection using convolutional neural networks,

    M. H. Yap, G. Pons, J. Marti, S. Ganau, M. Sentis, R. Zwiggelaar, A. K. Davison, and R. Marti, “Automated breast ultrasound lesions detection using convolutional neural networks,”IEEE journal of biomedical and health informatics, vol. 22, no. 4, pp. 1218–1226, 2017

    2017

  32. [40]

    Dataset of breast ultrasound images,

    W. Al-Dhabyani, M. Gomaa, H. Khaled, and A. Fahmy, “Dataset of breast ultrasound images,”Data in brief, vol. 28, p. 104863, 2020

    2020

  33. [41]

    Stu-hospital,

    xbhlk, “Stu-hospital,” https://github.com/xbhlk/STU-Hospital, 2024, ac- cessed: 2025-06-15

    2024

  34. [42]

    Segmentation and recognition of breast ultrasound images based on an expanded u-net,

    Y . Guo, X. Duan, C. Wang, and H. Guo, “Segmentation and recognition of breast ultrasound images based on an expanded u-net,”Plos one, vol. 16, no. 6, p. e0253202, 2021

    2021

  35. [43]

    Multi-task learning for thyroid nodule segmentation with thyroid region prior,

    H. Gong, G. Chen, R. Wang, X. Xie, M. Mao, Y . Yu, F. Chen, and G. Li, “Multi-task learning for thyroid nodule segmentation with thyroid region prior,” in2021 IEEE 18th international symposium on biomedical imaging (ISBI). IEEE, 2021, pp. 257–261

    2021

  36. [44]

    Less is more: Adaptive curriculum learning for thyroid nodule diagnosis,

    H. Gong, H. Cheng, Y . Xie, S. Tan, G. Chen, F. Chen, and G. Li, “Less is more: Adaptive curriculum learning for thyroid nodule diagnosis,” in International Conference on Medical Image Computing and Computer- Assisted Intervention. Springer, 2022, pp. 248–257

    2022

  37. [45]

    Algorithm guided outlining of 105 pancreatic cancer liver metastases in ultrasound,

    A. Hann, L. Bettac, M. M. Haenle, T. Graeter, A. W. Berger, J. Drey- haupt, D. Schmalstieg, W. G. Zoller, and J. Egger, “Algorithm guided outlining of 105 pancreatic cancer liver metastases in ultrasound,” Scientific Reports, vol. 7, no. 1, p. 12779, 2017

    2017

  38. [46]

    Improving artificial intelligence pipeline for liver malignancy diagnosis using ultrasound images and video frames,

    Y . Xu, B. Zheng, X. Liu, T. Wu, J. Ju, S. Wang, Y . Lian, H. Zhang, T. Liang, Y . Sanget al., “Improving artificial intelligence pipeline for liver malignancy diagnosis using ultrasound images and video frames,” Briefings in Bioinformatics, vol. 24, no. 1, p. bbac569, 2023

    2023

  39. [47]

    U2-bench: Benchmarking large vision-language models on ultrasound understanding,

    A. Le, H. Liu, Y . Wang, Z. Liu, R. Zhu, T. Weng, J. Yu, B. Wang, Y . Wu, K. Yanet al., “U2-bench: Benchmarking large vision-language models on ultrasound understanding,”arXiv preprint arXiv:2505.17779, 2025

    2025

  40. [48]

    Hamid, I

    S. Hamid, I. A. Donaldson, Y . Hu, R. Rodell, B. Villarini, E. Bonmati, P. Tranter, S. Punwani, H. S. Sidhu, S. Williset al., “The smarttarget biopsy trial: a prospective, within-person randomised, blinded trial comparing the accuracy of visual-registration and magnetic resona...

    2019

  41. [49]

    The open kidney ultrasound data set,

    R. Singla, C. Ringstrom, G. Hu, V . Lessoway, J. Reid, C. Nguan, and R. Rohling, “The open kidney ultrasound data set,” inInternational Workshop on Advances in Simplifying Medical Ultrasound. Springer, 2023, pp. 155–164

    2023

  42. [50]

    Bus-bra: a breast ultrasound dataset for assessing computer-aided diagnosis systems,

    W. G ´omez-Flores, M. J. Gregorio-Calas, and W. Coelho de Albu- querque Pereira, “Bus-bra: a breast ultrasound dataset for assessing computer-aided diagnosis systems,”Medical Physics, vol. 51, no. 4, pp. 3110–3123, 2024

    2024

  43. [51]

    Thyroid nodule segmentation and classification in ultrasound images,

    J. Zhou, X. Jia, D. Ni, A. Noble, R. Huang, T. Tan, and M. T. Van, “Thyroid nodule segmentation and classification in ultrasound images,” Mar. 2020. [Online]. Available: https://doi.org/10.5281/zenodo.3715942

    2020 doi

  44. [52]

    Luminous database: lumbar multifidus muscle segmentation from ultrasound images,

    C. J. Belasso, B. Behboodi, H. Benali, M. Boily, H. Rivaz, and M. Fortin, “Luminous database: lumbar multifidus muscle segmentation from ultrasound images,”BMC Musculoskeletal Disorders, vol. 21, pp. 1–11, 2020

    2020

  45. [53]

    Lora: Low-rank adaptation of large language models

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, W. Chenet al., “Lora: Low-rank adaptation of large language models.” ICLR, vol. 1, no. 2, p. 3, 2022

    2022

  46. [54]

    Generalized intersection over union: A metric and a loss for bounding box regression,

    H. Rezatofighi, N. Tsoi, J. Gwak, A. Sadeghian, I. Reid, and S. Savarese, “Generalized intersection over union: A metric and a loss for bounding box regression,” inProceedings of the IEEE/CVF conference on com- puter vision and pattern recognition, 2019, pp. 658–666

    2019

  47. [55]

    Focal loss for dense object detection,

    T.-Y . Lin, P. Goyal, R. Girshick, K. He, and P. Doll ´ar, “Focal loss for dense object detection,” inProceedings of the IEEE international conference on computer vision, 2017, pp. 2980–2988

    2017

  48. [56]

    The pascal visual object classes (voc) challenge,

    M. Everingham, L. Van Gool, C. K. Williams, J. Winn, and A. Zisser- man, “The pascal visual object classes (voc) challenge,”International journal of computer vision, vol. 88, pp. 303–338, 2010

    2010

  49. [57]

    Measures of the amount of ecologic association between species,

    L. R. Dice, “Measures of the amount of ecologic association between species,”Ecology, vol. 26, no. 3, pp. 297–302, 1945

    1945

  50. [58]

    Decoupled weight decay regularization,

    I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” arXiv preprint arXiv:1711.05101, 2017

    2017 arXiv

  51. [59]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Loet al., “Segment anything,” inProceedings of the IEEE/CVF international conference on computer vision, 2023, pp. 4015–4026

    2023

  52. [60]

    Biomedcoop: Learning to prompt for biomedical vision-language models,

    T. Koleilat, H. Asgariandehkordi, H. Rivaz, and Y . Xiao, “Biomedcoop: Learning to prompt for biomedical vision-language models,” inProceed- ings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 14 766–14 776

    2025

Pith tools

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