Pith. sign in

REVIEW 3 major objections 4 minor 45 references

TextSAM-EUS: Text Prompt Learning for SAM to Accurately Segment Pancreatic Tumor in Endoscopic Ultrasound

T0 review · 3 major / 4 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 Text prompts alone can drive SAM to 82.69% Dice on pancreatic EUS tumor segmentation, with no manual points or boxes at inference.

desk verdict A competent parameter-efficient text-driven SAM adaptation for EUS tumor segmentation, with a real but fixable evaluation flaw: the final hyperparameters are selected on the test set, and the significance tests are frame-level. read the letter →

arxiv 2507.18082 v3 pith:B5VFFZP5 submitted 2025-07-24 cs.CV cs.AI

Pascal Spiegler , Taha Koleilat , Arash Harirpoush , Corey S. Miller , Hassan Rivaz , Marta Kersten-Oertel , Yiming Xiao This is my paper · ORCID
classification cs.CVcs.AI
keywords textpromptlearningSegmentAnythingModelpancreatictumorsegmentationendoscopicultrasoundlow-rankadaptationBiomedCLIPcontextoptimizationparameter-efficientfine-tuning
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

TextSAM-EUS claims that a Segment Anything Model, adapted only through low-rank updates and learnable text tokens, can segment pancreatic tumors in endoscopic ultrasound without any manual geometric prompts at inference. On the public Endoscopic Ultrasound Database of the Pancreas, the fully automatic variant reports 82.69% Dice and 85.28% NSD, and the paper states this significantly outperforms all automatic baselines it compared against (p < 0.05) while tuning 0.86% of SAM's parameters. A variant given one ground-truth point and a perturbed box reports 83.10% Dice and 85.70% NSD, matching or exceeding MedSAM, which receives ground-truth boxes. The practical stake is that a clinician could obtain tumor delineation by specifying a class name, avoiding the radiological judgment that geometric prompting requires.

What carries the argument

The load-bearing object is the learned text prompt rather than a geometric prompt: a small set of context tokens (b = 4) concatenated with the class name 'tumor' and optimized through the first t = 12 Transformer layers of the BiomedCLIP text encoder, then mapped by an MLP adapter into SAM's prompt encoder. LoRA with rank 16 on the SAM image encoder and mask decoder makes the base model trainable with only 1.69 million parameters, about 0.86% of SAM's total. The mask decoder fuses the text-derived tokens with dense image embeddings, and an iterative refinement step adds a bounding box and centroid extracted from the initial prediction, giving the final 82.69% Dice automatic result.

What would settle it

Freeze the final configuration (LoRA rank 16, 4 context tokens, depth 12, box plus centroid refinement) using only the training and validation splits, then evaluate it once on a test set that has never influenced any choice. If the automatic variant's Dice drops below AutoSAM or AutoSAMUS, or the p < 0.05 superiority claim reverses, the central claim is refuted. A second check is to report per-patient confidence intervals to see whether the roughly 1.4–1.65% Dice gap is stable across patients.

Watch

Extended reading notes

Core claim

The central discovery is that language can replace geometry as the prompt modality for SAM in a noisy, low-contrast medical imaging setting. TextSAM-EUS threads learnable context tokens through the first 12 layers of the BiomedCLIP text encoder, projects the resulting embedding into SAM's prompt-token space with a two-layer MLP adapter, and feeds those text tokens to SAM's mask decoder alongside LoRA-adapted image features. After an initial text-only mask, it extracts the predicted mask's bounding box and centroid and feeds them back for two refinement iterations. The paper's headline assertion is that this automatic, text-driven pipeline reaches 82.69% Dice and 85.28% NSD on the public pancreatic EUS dataset, significantly outperforming all automatic baselines it compares against (p < 0.05), and that with manual geometric prompts it reaches 83.10% Dice and 85.70% NSD, matching or exceeding the strongest prompt-based baseline, MedSAM. The paper also claims this is the first use of prompt learning in SAM-based medical segmentation, and that the success shows text-to-ViT feature alignment carries more value for EUS than adding an external CNN encoder path.

Load-bearing premise

The load-bearing premise is that the held-out test set was not used to choose the final configuration; the paper reports ablation results on that same test set, so if those design choices were made on it, the 82.69% Dice and the significance comparisons are optimistically biased.

Editorial extensions

If this is right

  • If the claim holds, automatic EUS tumor segmentation no longer requires a human to place points or boxes; the same frozen SAM can be redirected to a new target by changing the learned text prompt and fine-tuning a very small parameter set.
  • The parameter footprint of 1.69 million trainable parameters means the method could run on clinical hardware with a single GPU, a different regime from full SAM fine-tuning or training a U-Net from scratch.
  • Text-only prompting already reaches 82.00% Dice before refinement, and manual geometric prompts add only about one point of Dice, suggesting language guidance carries most of the discriminative information needed for this binary task.
  • The reported significant superiority over AutoSAM and AutoSAMUS (p < 0.05) implies that, for EUS, learning text-to-feature alignment is a better use of adaptation capacity than learning automatic point or box generators.

Reading between the lines

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

  • Because the text encoder is frozen except for the context tokens, a testable extension is to check whether the same learned prompt tokens transfer across EUS scanners or patients; if they do, adapting to a new lesion type could reduce to retraining only the context tokens.
  • The comparison to MedSAM is asymmetric, since MedSAM receives ground-truth boxes while TextSAM-EUS receives none; an editorial consequence is that a head-to-head with MedSAM using identical predicted boxes would clarify whether text guidance or box guidance is the stronger prior.
  • A stricter evaluation protocol that holds out a separate fold for hyperparameter selection could reveal whether the reported 1.4–1.65% Dice gap over AutoSAM and AutoSAMUS persists; this is directly testable with the same public dataset.
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

3 major / 4 minor

Summary. The paper proposes TextSAM-EUS, a parameter-efficient adaptation of the Segment Anything Model (SAM) for automatic pancreatic tumor segmentation in endoscopic ultrasound (EUS). The method combines LoRA-based fine-tuning of SAM's image encoder and mask decoder with learnable text prompt tokens processed through a frozen BiomedCLIP text encoder, replacing manual geometric prompts with learned text-driven prompts at inference. An iterative refinement stage derives box and centroid prompts from the initial text-driven mask and feeds them back into SAM. On the public Endoscopic Ultrasound Database of the Pancreas, the authors report 82.69% Dice and 85.28% NSD for the fully automatic variant, and 83.10% Dice and 85.70% NSD when manual prompts are supplied, while tuning only 0.86% (1.69M) of SAM's parameters. Ablations study LoRA rank, context-token length, prompt-injection depth, and refinement-prompt choices. The paper claims significant improvements over all automatic baselines and over several manual-prompted SAM variants.

Significance. If the reported evaluation were unbiased, the contribution would be practically valuable: a text-driven, prompt-free SAM adaptation for a noisy, low-contrast ultrasound modality, with a very small trainable parameter footprint and a clear baseline comparison against both fully supervised and foundation-model methods. The code release and use of a public dataset are strengths that support reproducibility. The central novelty, integrating CoOp-style prompt learning with SAM via a biomedical text encoder, is plausible and timely. However, the evaluation protocol contains a load-bearing flaw: the held-out test set is used for hyperparameter and design selection, and the statistical comparisons ignore within-patient correlation. These issues directly affect the headline accuracy numbers and the 'significant outperformance' claim, so the current evidence does not yet support the stated state-of-the-art conclusion.

major comments (3)
  1. [Section 4.5, Tables 3-6] All four ablations (LoRA rank, context-token length, prompt-injection depth, and refinement-prompt choice) are explicitly evaluated on the held-out test set: 'All results are reported on the held-out segmentation test set.' The final configuration (r=16, b=4, t=12, box+centroid) is then the best-scoring combination on that same test set, and the reported 82.69% Dice and 85.28% NSD come from that configuration. This makes the headline numbers selection-optimized estimates, not unbiased predictions of performance on the test distribution. The problem is compounded in the comparison with baselines: baseline hyperparameters follow the recommendations of their original papers, while TextSAM-EUS's hyperparameters are selected on the test set. The paper should freeze all hyperparameters using only the validation split (or use nested cross-validation) and report test metrics for the frozen configuration; Section 4.3's statement that 'model selection is based on the lowest validation loss' does not resolve this, because the ablation choices are not made on validation data.
  2. [Section 4.4, statistical analysis] The significance claims are based on two-tailed paired-sample t-tests over 4,185 frames drawn from only 4 test patients. Treating each frame as an independent observation ignores the strong within-patient correlation of EUS frames, so the effective sample size for the comparison is much closer to 4 than to 4,185, and the reported p-values are anti-conservatively biased. The paper should report patient-level metrics (e.g., per-patient Dice and NSD) and use a patient-level or cluster-robust significance test, with appropriate correction for the multiple pairwise comparisons against the many baselines. Without this, the statement that TextSAM-EUS 'significantly outperforms all automatic baselines' is not supported.
  3. [Section 4.4, MedSAM comparison] The text says the automatic TextSAM-EUS and MedSAM yield 'nearly identical DSC (82.69 vs. 82.66; p = 0.906)', but then claims to 'significantly outperform all automatic baselines'. MedSAM is not automatic, since it is supplied with ground-truth bounding boxes, so the claim is internally consistent only under the paper's definition of 'automatic'. However, the comparison also shows that MedSAM achieves significantly higher NSD (85.75 vs. 85.28; p = 0.035). The Discussion should state this discrepancy clearly rather than describing the two as nearly identical across both metrics; the current wording overstates the closeness of the result.
minor comments (4)
  1. [Section 5] The preliminary ablation in which 'swapping the SAMUS encoder into our framework yielded a maximum of 67.53% DSC' is mentioned only in prose and not reported in any table or with experimental details; either add the full setup and results to Section 4.5 or remove the claim.
  2. [Abstract and Introduction] The phrase 'the first attempt to incorporate prompt learning in SAM-based medical image segmentation' is strong given the authors' own MedCLIP-SAM family [23,24] and the existence of text-guided SAM variants such as CC-SAM [10]; the novelty claim should be qualified to specify what exactly is new relative to these prior text-driven SAM adaptations.
  3. [Section 4.3] The description of the prompt-injection depth is ambiguous: 'a depth of 12 layers (threaded through the first 12 Transformer layers)' could be read as two different quantities. Clarify that BiomedCLIP has 12 blocks total and that t=12 means the learnable tokens are updated through all blocks.
  4. [Table 2] SwinUNet's NSD of 10.22 ± 6.64 with a Dice of 59.69 is an extreme outlier relative to its DSC; a brief explanation of this behavior (e.g., boundary tolerance effects) would help the reader interpret the table.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the reported segmentation metrics are empirical measurements against a public dataset, and the method's components are defined independently of the target result.

full rationale

I find no circular derivation in TextSAM-EUS. The central claim—that learned text prompts plus LoRA-adapted SAM reach 82.69% DSC and 85.28% NSD on the Endoscopic Ultrasound Database of the Pancreas—is an empirical result obtained by training on a patient-disjoint training set and evaluating on a held-out test set; the Dice score is not an input to the loss or to the architecture. The text-driven pathway is specified by Eqs. (5)-(12) as a composition of the externally pretrained BiomedCLIP text encoder, LoRA updates, and a trained MLP adapter; no equation defines the reported metric in terms of itself. Same-group citations (MedCLIP-SAM, MedCLIP-SAMv2, BiomedCoOp) appear only as related-work context and motivation, not as evidence that the proposed model performs well, and the head-to-head comparisons are against independently implemented or published baselines. The iterative refinement module uses the model's own predicted mask to derive box and centroid prompts; this is a self-contained post-processing step rather than a fitted input masquerading as a prediction. The ablations in Sec. 4.5 are evaluated on the held-out test set, which raises a legitimate statistical-validity concern about model selection and the unbiasedness of the headline number, but selecting among configurations using test performance is not circular reasoning: it does not make the derived quantity equivalent to its inputs. I cannot exhibit any equation or self-citation chain that reduces the claim to its own assumptions, so the appropriate finding is no significant circularity.

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

The central empirical result rests on standard supervised learning assumptions and on the correctness of expert annotations, plus the paper's choice to treat the test set as a design set during ablations. No new mathematical axioms or physical entities are introduced.

free parameters (5)
  • LoRA rank r = 16
    Selected from ablation on the held-out test set (Table 3); rank 16 gives 82.69% Dice while rank 4 gives 81.33% and rank 1 collapses to 44.00%.
  • Number of learnable context tokens b = 4
    Ablation in Table 4 on the test set; 4 tokens are best, 8 tokens degrade to 66.74% Dice, 12 tokens recover to 82.10%.
  • Prompt injection depth t = 12
    Ablation in Table 5 on the test set; deeper injection improves Dice from 78.32% at t=1 to 82.69% at t=12.
  • Geometric refinement prompt combination = bounding box + centroid point
    Ablation in Table 6 on the test set; box plus centroid gives the highest Dice and NSD over other prompt combinations.
  • Iterative refinement iterations = 2
    Stated in Section 3.5 as the number of refinement iterations; no ablation on iteration count is reported.
assumptions (4)
  • domain assumption Expert pixel-level annotations in the EUS dataset are accurate ground truth.
    All DSC and NSD numbers depend on these labels; no inter-observer variability is reported in Section 4.1.
  • domain assumption The 4,185 frames from 4 test patients can be treated as independent samples for paired t-tests.
    Section 4.4 reports p-values from two-tailed paired-sample t-tests across all test cases; frames from the same patient are correlated, so these p-values may overstate significance.
  • ad hoc to paper Choosing hyperparameters and refinement design on the held-out test set still yields an unbiased estimate of final performance.
    Section 4.5 reports all ablations on the test set and the best configuration is selected from those results; this assumes that test-set-based design does not bias the reported final metrics.
  • domain assumption BiomedCLIP text embeddings, after projection, provide useful semantic conditioning for SAM's mask decoder.
    The method depends on the adapter mapping text embeddings into SAM prompt space; this is supported only empirically on the single EUS dataset used here.

how reviews work

0 comments
Cite this review

Pith. "Pith review of TextSAM-EUS: Text Prompt Learning for SAM to Accurately Segment Pancreatic Tumor in Endoscopic Ultrasound." pith.science (2026). https://pith.science/paper/B5VFFZP5

@misc{pith2026250718082,
  author       = {Pith},
  title        = {Pith review of: TextSAM-EUS: Text Prompt Learning for SAM to Accurately Segment Pancreatic Tumor in Endoscopic Ultrasound},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B5VFFZP5}},
  note         = {Machine review of arXiv:2507.18082}
}
read the original abstract

Pancreatic cancer carries a poor prognosis and relies on endoscopic ultrasound (EUS) for targeted biopsy and radiotherapy. However, the speckle noise, low contrast, and unintuitive appearance of EUS make segmentation of pancreatic tumors with fully supervised deep learning (DL) models both error-prone and dependent on large, expert-curated annotation datasets. To address these challenges, we present TextSAM-EUS, a novel, lightweight, text-driven adaptation of the Segment Anything Model (SAM) that requires no manual geometric prompts at inference. Our approach leverages text prompt learning (context optimization) through the BiomedCLIP text encoder in conjunction with a LoRA-based adaptation of SAM's architecture to enable automatic pancreatic tumor segmentation in EUS, tuning only 0.86% of the total parameters. On the public Endoscopic Ultrasound Database of the Pancreas, TextSAM-EUS with automatic prompts attains 82.69% Dice and 85.28% normalized surface distance (NSD), and with manual geometric prompts reaches 83.10% Dice and 85.70% NSD, outperforming both existing state-of-the-art (SOTA) supervised DL models and foundation models (e.g., SAM and its variants). As the first attempt to incorporate prompt learning in SAM-based medical image segmentation, TextSAM-EUS offers a practical option for efficient and robust automatic EUS segmentation. Code is available at https://github.com/HealthX-Lab/TextSAM-EUS .

Figures

Figures reproduced from arXiv: 2507.18082 by the authors.

Figure 1
Figure 1. Overview of the TEXTSAM-EUS framework. We finetune context tokens in the BiomedCLIP text encoder Et to produce text embeddings T, which are projected via an adapter E text prompt into the SAM prompt token space. These guide the SAM mask decoder Dmask, which fuses them with image features F = Eimg(X) from the ViT-based image encoder Eimg, adapted using LoRA. An iterative segmentation refinement step incorporates geom… view at source ↗
Figure 2
Figure 2. Qualitative comparison on a representative EUS slice, comparing the most competitive nine baselines alongside T [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 25 canonical work pages

  1. [1]

    Making the most of text semantics to improve biomedical vision–language processing

    Benedikt Boecking, Naoto Usuyama, Shruthi Bannur, Daniel C Castro, Anton Schwaighofer, Stephanie Hyland, Maria Wetscherek, Tristan Naumann, Aditya Nori, Javier Alvarez-Valle, et al. Making the most of text semantics to improve biomedical vision–language processing. In Euro- pean conference on computer vision , pages 1–21. Springer,

  2. [2]

    Global cancer statistics 2022: Globocan estimates of incidence and mortality worldwide for 36 can- cers in 185 countries

    Freddie Bray, Mathieu Laversanne, Hyuna Sung, Jacques Ferlay, Rebecca L Siegel, Isabelle Soerjomataram, and Ahmedin Jemal. Global cancer statistics 2022: Globocan estimates of incidence and mortality worldwide for 36 can- cers in 185 countries. CA: a cancer journal for clinicians , 74(3):229–263, 2024. 1

    work page 2022

  3. [3]

    Sabuncu, John Guttag, and Adrian V

    Victor Ion Butoi, Jose Javier Gonzalez Ortiz, Tianyu Ma, Mert R. Sabuncu, John Guttag, and Adrian V . Dalca. Uni- verseg: Universal medical image segmentation, 2023. 5

    work page 2023

  4. [4]

    Swin-unet: Unet-like pure transformer for medical image segmentation

    Hu Cao, Yueyue Wang, Joy Chen, Dongsheng Jiang, Xi- aopeng Zhang, Qi Tian, and Manning Wang. Swin-unet: Unet-like pure transformer for medical image segmentation. In European conference on computer vision, pages 205–218. Springer, 2022. 5

    work page 2022

  5. [5]

    Domain-controlled prompt learning

    Qinglong Cao, Zhengqin Xu, Yuntian Chen, Chao Ma, and Xiaokang Yang. Domain-controlled prompt learning. InPro- ceedings of the AAAI Conference on Artificial Intelligence , pages 936–944, 2024. 3

    work page 2024

  6. [6]

    Sam on medical images: A comprehensive study on three prompt modes

    Dongjie Cheng, Ziyuan Qin, Zekun Jiang, Shaoting Zhang, Qicheng Lao, and Kang Li. Sam on medical images: A comprehensive study on three prompt modes. ArXiv, abs/2305.00035, 2023. 2

    arXiv 2023

  7. [7]

    Does clip benefit visual question answering in the medical domain as much as it does in the general domain?, 2021

    Sedigheh Eslami, Gerard de Melo, and Christoph Meinel. Does clip benefit visual question answering in the medical domain as much as it does in the general domain?, 2021. 3

    2021

  8. [8]

    Clip-adapter: Better vision-language models with feature adapters

    Peng Gao, Shijie Geng, Renrui Zhang, Teli Ma, Rongyao Fang, Yongfeng Zhang, Hongsheng Li, and Yu Qiao. Clip-adapter: Better vision-language models with feature adapters. International Journal of Computer Vision, 132(2): 581–595, 2024. 3

    2024

Show all 45 references
  1. [9]

    3dsam-adapter: Holistic adaptation of sam from 2d to 3d for promptable medical image segmentation, 2023

    Shizhan Gong, Yuan Zhong, Wenao Ma, Jinpeng Li, Zhao Wang, Jingyang Zhang, Pheng-Ann Heng, and Qi Dou. 3dsam-adapter: Holistic adaptation of sam from 2d to 3d for promptable medical image segmentation, 2023. 2

    2023

  2. [10]

    Cc-sam: Sam with cross-feature attention and context for ultrasound image seg- mentation

    Shreyank N Gowda and David A Clifton. Cc-sam: Sam with cross-feature attention and context for ultrasound image seg- mentation. In European Conference on Computer Vision , pages 108–124. Springer, 2024. 3

    2024

  3. [11]

    Domain-specific language model pre- training for biomedical natural language processing

    Yu Gu, Robert Tinn, Hao Cheng, Michael Lucas, Naoto Usuyama, Xiaodong Liu, Tristan Naumann, Jianfeng Gao, and Hoifung Poon. Domain-specific language model pre- training for biomedical natural language processing. ACM Transactions on Computing for Healthcare (HEALTH), 3(1): 1–2...

    2021

  4. [12]

    Lora: Low- rank adaptation of large language models

    Edward J Hu, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. Lora: Low- rank adaptation of large language models. In International Conference on Learning Representations, 2021. 4

    2021

  5. [13]

    Cascade segmentation framework for pancreatic tumor in ultrasound images

    Xin Huang, Jin Wang, Yuan Zhang, and Haibo Liu. Cascade segmentation framework for pancreatic tumor in ultrasound images. IEEE Access, 8:126806–126815, 2020. 2

    2020

  6. [14]

    Push the boundary of sam: A pseudo-label correction framework for medical segmenta- tion, 2023

    Ziyi Huang, Hongshan Liu, Haofeng Zhang, Xueshen Li, Haozhe Liu, Fuyong Xing, Andrew Laine, Elsa Angelini, Christine Hendon, and Yu Gan. Push the boundary of sam: A pseudo-label correction framework for medical segmenta- tion, 2023. 2

    2023

  7. [15]

    nnu-net: a self-configuring method for deep learning-based biomedical image segmen- tation

    Fabian Isensee, Paul F Jaeger, Simon AA Kohl, Jens Pe- tersen, and Klaus H Maier-Hein. nnu-net: a self-configuring method for deep learning-based biomedical image segmen- tation. Nature methods, 18(2):203–211, 2021. 5

    2021

  8. [16]

    Endoscopic ultrasound database of the pancreas

    Mar ´ıa Jaramillo, Josu ´e Ruano, Mart´ın G´omez, and Eduardo Romero. Endoscopic ultrasound database of the pancreas. In 16th International Symposium on Medical Information Pro- cessing and Analysis, pages 130–135. SPIE, 2020. 5

    2020

  9. [17]

    Scaling up visual and vision-language representa- tion learning with noisy text supervision

    Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc Le, Yun-Hsuan Sung, Zhen Li, and Tom Duerig. Scaling up visual and vision-language representa- tion learning with noisy text supervision. In International conference on machine learning, pages 4904–4916. PMLR,

  10. [18]

    Maple: Multi-modal prompt learning

    Muhammad Uzair Khattak, Hanoona Rasheed, Muhammad Maaz, Salman Khan, and Fahad Shahbaz Khan. Maple: Multi-modal prompt learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19113–19122, 2023. 3

    2023

  11. [19]

    Self-regulating prompts: Foundational model adaptation without forgetting

    Muhammad Uzair Khattak, Syed Talal Wasim, Muzam- mal Naseer, Salman Khan, Ming-Hsuan Yang, and Fa- had Shahbaz Khan. Self-regulating prompts: Foundational model adaptation without forgetting. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 15...

    2023

  12. [20]

    Learning to prompt with text only supervision for vision- language models

    Muhammad Uzair Khattak, Muhammad Ferjad Naeem, Muzammal Naseer, Luc Van Gool, and Federico Tombari. Learning to prompt with text only supervision for vision- language models. arXiv preprint arXiv:2401.02418 , 2024. 4

    2024 arXiv

  13. [21]

    Berg, Wan-Yen Lo, Piotr Doll ´ar, and Ross Girshick

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer White- head, Alexander C. Berg, Wan-Yen Lo, Piotr Doll ´ar, and Ross Girshick. Segment anything, 2023. 1, 2, 3

    2023

  14. [22]

    Transformer les soins de sant ´e gr ˆace `a des mod`eles de langage multimodaux universels, efficaces et ´evolutifs pour l’analyse biom ´edicale

    Taha Koleilat. Transformer les soins de sant ´e gr ˆace `a des mod`eles de langage multimodaux universels, efficaces et ´evolutifs pour l’analyse biom ´edicale. https://doi. org/10.69777/363874, 2025. Bourses de doctorat en recherche, num´ero de demande 363874, Fonds de recher...

    2025 doi

  15. [23]

    Medclip-sam: Bridging text and image towards universal medical image segmentation

    Taha Koleilat, Hojat Asgariandehkordi, Hassan Rivaz, and Yiming Xiao. Medclip-sam: Bridging text and image towards universal medical image segmentation. In In- ternational Conference on Medical Image Computing and Computer-Assisted Intervention , pages 643–653. Springer,

  16. [24]

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

    Taha Koleilat, Hojat Asgariandehkordi, Hassan Rivaz, and Yiming Xiao. Medclip-samv2: Towards universal text-driven medical image segmentation. arXiv preprint arXiv:2409.19483, 2024. 2, 3

    2024 arXiv

  17. [25]

    Biomedcoop: Learning to prompt for biomedi- cal vision-language models

    Taha Koleilat, Hojat Asgariandehkordi, Hassan Rivaz, and Yiming Xiao. Biomedcoop: Learning to prompt for biomedi- cal vision-language models. In Proceedings of the Computer Vision and Pattern Recognition Conference , pages 14766– 14776, 2025. 3

    2025

  18. [26]

    Be- yond adapting sam: Towards end-to-end ultrasound image segmentation via auto prompting

    Xian Lin, Yangyang Xiang, Li Yu, and Zengqiang Yan. Be- yond adapting sam: Towards end-to-end ultrasound image segmentation via auto prompting. In International Confer- ence on Medical Image Computing and Computer-Assisted Intervention, pages 24–34. Springer, 2024. 2, 3, 4, 5

    2024

  19. [27]

    Semi-supervised segmentation of pancreatic tumor in ul- trasound images with multi-task consistency learning

    Yifan Liu, Lanfen Jin, Minhao Xie, Kai Wang, and Xiahai Xu. Semi-supervised segmentation of pancreatic tumor in ul- trasound images with multi-task consistency learning. Med- ical Image Analysis, 70:101998, 2021. 2

    2021

  20. [28]

    Decoupled weight decay regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. 6

    2017 arXiv

  21. [29]

    A deep learning framework for pancreatic tumor segmentation in ultrasound images

    Yu Lu, Jin Xue, Minhao Xie, Kai Wang, and Xiahai Xu. A deep learning framework for pancreatic tumor segmentation in ultrasound images. Ultrasound in Medicine & Biology, 47 (4):968–982, 2021. 2

    2021

  22. [30]

    Segment anything in medical images

    Jun Ma, Yuting He, Feifei Li, Lin Han, Chenyu You, and Bo Wang. Segment anything in medical images. Nature Communications, 15(1), 2024. 2, 5

    2024

  23. [31]

    Feasibility and safety of endoscopic ultrasound-guided dif- fusing alpha emitter radiation therapy for advanced pan- creatic cancer: Preliminary data

    Corey S Miller, Magali Lecavalier-Barsoum, Kim Ma, Miriam Santos Dutra, Youri Kaitoukov, Boris Bahoric, Nada Tomic, Francine Dinelle, Shirin Enger, Gerald Batist, et al. Feasibility and safety of endoscopic ultrasound-guided dif- fusing alpha emitter radiation therapy for adva...

    2024

  24. [32]

    Learning transferable visual models from natural language supervision, 2021

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. Learning transferable visual models from natural language supervision, 2021. 3

    2021

  25. [33]

    Autosam: Adapting sam to medical images by overloading the prompt encoder

    Tal Shaharabany, Aviad Dahan, Raja Giryes, and Lior Wolf. Autosam: Adapting sam to medical images by overloading the prompt encoder. arXiv preprint arXiv:2306.06370, 2023. 2, 5

    2023 arXiv

  26. [34]

    Planification automatique de la trajectoire des ´Electrodes pour la chirurgie de stimulation c´er´ebrale pro- fonde connectomique

    Pascal Spiegler. Planification automatique de la trajectoire des ´Electrodes pour la chirurgie de stimulation c´er´ebrale pro- fonde connectomique. https://doi.org/10.69777/ 348625, 2025. Bourse de maˆıtrise en recherche, num´ero de demande 348625, Fonds de recherche du Qu´ebe...

    2025

  27. [35]

    Weakly supervised intracranial hemorrhage segmentation with yolo and an uncertainty rectified segment anything model

    Pascal Spiegler, Amirhossein Rasoulian, and Yiming Xiao. Weakly supervised intracranial hemorrhage segmentation with yolo and an uncertainty rectified segment anything model. In MICCAI Challenge on Ischemic Stroke Lesion Segmentation, pages 12–21. Springer, 2024. 2

    2024

  28. [36]

    Self-prompting large vision models for few-shot medical image segmenta- tion

    Qi Wu, Yuyao Zhang, and Marawan Elbatel. Self-prompting large vision models for few-shot medical image segmenta- tion. In MICCAI workshop on domain adaptation and rep- resentation transfer, pages 156–167. Springer, 2023. 2, 5

    2023

  29. [37]

    Am ´eliorer le prognostic neurochirurgical avec l’imagerie multimodale et `a la connectivit ´e c ´er´ebrale

    Yiming Xiao. Am ´eliorer le prognostic neurochirurgical avec l’imagerie multimodale et `a la connectivit ´e c ´er´ebrale. https : / / doi . org / 10 . 69777 / 330745, 2024. Chercheurs-boursiers, num ´ero de demande 330745, Fonds de recherche du Qu´ebec (FRQ). 9

    2024

  30. [38]

    Advances in medical image seg- mentation: A comprehensive review of traditional, deep learning and hybrid approaches

    Yan Xu, Rixiang Quan, Weiting Xu, Yi Huang, Xiaolong Chen, and Fengyuan Liu. Advances in medical image seg- mentation: A comprehensive review of traditional, deep learning and hybrid approaches. Bioengineering, 11(10): 1034, 2024. 3

    2024

  31. [39]

    Customized segment any- thing model for medical image segmentation

    Kaidong Zhang and Dong Liu. Customized segment any- thing model for medical image segmentation. arXiv preprint arXiv:2304.13785, 2023. 2, 5, 6

    2023 arXiv

  32. [40]

    Tip-adapter: Training-free clip-adapter for better vision- language modeling

    Renrui Zhang, Rongyao Fang, Wei Zhang, Peng Gao, Kunchang Li, Jifeng Dai, Yu Qiao, and Hongsheng Li. Tip-adapter: Training-free clip-adapter for better vision- language modeling. arXiv preprint arXiv:2111.03930, 2021. 3

    2021 arXiv

  33. [41]

    Lungren, Tristan Naumann, Sheng Wang, and Hoifung Poon

    Sheng Zhang, Yanbo Xu, Naoto Usuyama, Hanwen Xu, Jaspreet Bagga, Robert Tinn, Sam Preston, Rajesh Rao, Mu Wei, Naveen Valluri, Cliff Wong, Andrea Tupini, Yu Wang, Matt Mazzola, Swadheen Shukla, Lars Liden, Jian- feng Gao, Matthew P. Lungren, Tristan Naumann, Sheng Wang, and Ho...

    2024

  34. [42]

    Biomed- parse: a biomedical foundation model for image pars- ing of everything everywhere all at once

    Theodore Zhao, Yu Gu, Jianwei Yang, Naoto Usuyama, Ho Hin Lee, Tristan Naumann, Jianfeng Gao, Angela Crab- tree, Jacob Abel, Christine Moung-Wen, et al. Biomed- parse: a biomedical foundation model for image pars- ing of everything everywhere all at once. arXiv preprint arXiv:...

    2024 arXiv

  35. [43]

    Clip in medical imaging: A comprehensive sur- vey

    Zihao Zhao, Yuxiao Liu, Han Wu, Mei Wang, Yonghao Li, Sheng Wang, Lin Teng, Disheng Liu, Zhiming Cui, Qian Wang, et al. Clip in medical imaging: A comprehensive sur- vey. arXiv preprint arXiv:2312.07353, 2023. 3

    2023 arXiv

  36. [44]

    Conditional prompt learning for vision-language models

    Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Zi- wei Liu. Conditional prompt learning for vision-language models. In CVPR, pages 16816–16825, 2022. 3

    2022

  37. [45]

    Learning to prompt for vision-language models

    Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. Learning to prompt for vision-language models. IJCV, 130(9):2337–2348, 2022. 3

    2022

Pith tools

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