Pith. sign in

REVIEW 4 major objections 5 minor 54 references

Bootstrapping Vision-Language Model for Hysteroscopic Surgical Scene Segmentation

T0 review · 4 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read By bootstrapping the MedCLIP vision-language model with category-specific text prompts and a masked distillation branch, VLM-hyster achieves the best reported segmentation accuracy on hysteroscopic surgical scenes, with an overall IoU of…

desk verdict VLM-hyster is a solid applied paper with a genuinely new multicentre dataset and strong held-out numbers, but the text-guidance mechanism lacks a random-masking control and the external/prospective validations omit the strongest baselines. read the letter →

arxiv 2608.09302 v2 pith:FUFV4Z3E submitted 2026-08-10 cs.CV

classification cs.CV
keywords hysteroscopicsurgicalscenesegmentationvision-languagemodelmedicalimagemaskeddistillationtext-guidedcorrelationfilteringMedCLIPSAM2labelpropagationmulticentredataset
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 introduces VLM-hyster, a vision-language segmentation model for hysteroscopic surgical scenes, and argues that it produces the best reported pixel-wise masks for 15 categories of lesions and instruments. The method builds on the pretrained MedCLIP image encoder and a transformer decoder, then adds a masked distillation branch in which category-specific text prompts select and mask the visual tokens least correlated with the prompt, forcing the decoder to maintain its segmentation from the remaining tokens. On the paper's new held-out test set it reports DSC 82.51, MIoU 76.04, and OIoU 80.35, ahead of strong baselines such as Med-SAM (OIoU 74.29) and Med-VLM (OIoU 76.83). The paper also contributes a new multicentre dataset of 4,020 hysteroscopic images, with masks generated largely by SAM2 video propagation and verified by gynecologists, plus multicentre, prospective, and expert-assessed validation. If the claims hold, the method offers a practical route to real-time AI-assisted localization of visually similar lesions during hysteroscopy.

What carries the argument

The load-bearing mechanism is the masked distillation branch with text-guided correlation filtering. For each category, the text prompt 'the [category] in hysteroscopic image.' is encoded by MedCLIP's text encoder, and the [CLS] embedding is dot-producted with the image tokens of the encoder's last layer to yield a per-token correlation score. The branch then selects the 50% of tokens with the lowest correlation scores, randomly masks 25% of those selected tokens (an overall 12.5% masking ratio), and feeds the remaining tokens into a second decoder that shares the segmentation backbone's weights but drops its skip connections. A dice-style distillation loss, added with weight λ=0.75 to the main dice loss, forces the masked-view prediction to match the backbone's full prediction, teaching the model to lean on text-correlated visual evidence. Multi-scale features from encoder layers 4, 8, and 12 are passed into the main decoder so both low-level texture and high-level semantic cues are preserved.

What would settle it

Independently re-annotate every fifth frame of the full held-out test set by hand and recompute DSC, MIoU, and OIoU against the fresh manual masks; if the numbers fall materially, or if VLM-hyster's errors concentrate exactly where the propagated masks are wrong, the state-of-the-art claim would not survive. A quicker check is to extend the authors' 10-video validation (average IoU 91.1%) to the test videos and confirm that the auto-propagated masks stay above that level.

Watch

Extended reading notes

Core claim

The central discovery is that a text-prompted masked distillation branch can make a medical vision-language model decisively better at surgical scene segmentation. Across the 15 categories, VLM-hyster's image decoder is trained alongside a second decoder (shared weights, no skip connections) that receives a partially masked view of the image: the tokens dropped are those with the lowest dot-product correlation to the text prompt of a randomly chosen category, so the model must reproduce the full segmentation while only the text-correlated visual evidence survives. The authors report this mechanism raises DSC, MIoU, and OIoU by roughly 7 points over the same model without the branch (82.51 vs 75.13 DSC), and the resulting model outperforms all compared SAM-based and VLM-based methods on the held-out set, with improvements over Med-VLM significant at p<0.001.

Load-bearing premise

The accuracy numbers assume that the auto-propagated ground-truth masks produced by SAM2 from only five manually labeled frames per video are accurate enough that correcting the 8.3% of masks with overlap below 80% removes any systematic label bias that could inflate the reported scores.

Editorial extensions

If this is right

  • If the reported results replicate, VLM-hyster gives surgeons a real-time (15.8 FPS on a single A40) tool to localize 15 categories of lesions and instruments during hysteroscopy, including classes with nearly identical appearances such as endometrial polyps and polypoid hyperplasia.
  • Because the branch is driven only by text prompts, adding a new anatomical or instrument category requires writing a new prompt rather than retraining the backbone from scratch; the GPT-generated long-prompt variant (82.94 DSC with token filtering) suggests richer descriptions can further improve the model.
  • The multicentre and prospective validations (e.g., DSC 80.27 on a prospectively collected set) indicate the model transfers across hospitals and over time without additional training, which is the condition that matters for clinical deployment.
  • The new 4,020-image, 15-class dataset established by the paper gives the field a common benchmark that future hysteroscopic segmentation work can be measured against.

Reading between the lines

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

  • The same masked-distillation recipe may transfer to other narrow medical domains (e.g., cystoscopy, arthroscopy, or laparoscopy) that have no dedicated VLM yet, because the mechanism only needs a pretrained medical image-text encoder and a set of category prompts.
  • The paper's own prompt ablation suggests the bottleneck is not prompt wording but alignment noise: the detailed P3 prompt only helped after a textual-token filter was added, pointing to token-level cross-modal alignment as the key scaling lever for VLM-based segmentation.
  • A fair stress test of the claimed advantage would compare VLM-hyster against Med-VLM on a fully manually annotated subset; if the gap persists there, the masked distillation branch—not the SAM2 auto-annotation pipeline—is what carries the performance.
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

4 major / 5 minor

Summary. The paper introduces VLM-hyster, a vision-language-model-based method for hysteroscopic surgical scene segmentation. The model uses the MedCLIP image encoder as the backbone, a transformer decoder with multi-scale skip connections, and a masked distillation branch that filters visual tokens by correlation with category-specific text prompts during training. The authors construct a multicentric dataset of 4,020 hysteroscopic images with 15 categories, of which five frames per video are manually annotated and the remainder are propagated with SAM2 and partially corrected. On the held-out TJ-HS test set, VLM-hyster reports DSC 82.51, MIoU 76.04, and OIoU 80.35, outperforming CNN, SAM-based, and VLM-based baselines with paired t-tests and confidence intervals. The paper also reports multi-category experiments, per-category results, gynecologist assessments, multicentre and prospective validation, ablations over the masked distillation branch, masking ratio, loss weight, encoders, and text prompts, and a computational cost analysis.

Significance. If the results hold, the paper makes a useful contribution: it is the first VLM-based method for hysteroscopic scene segmentation, it provides a new multicentric dataset with 15 categories, and its strongest held-out comparisons are more extensive than typical for this area. The paper also reports confidence intervals and significance tests, includes ablation studies over several design choices, and releases code, which are all strengths. The main claims are plausible, but the central mechanism—text-guided masking—is not yet properly isolated, and the label-generation pipeline leaves residual uncertainty in the reported numbers. The contribution is likely to be of interest to the surgical vision and medical image analysis communities, provided the mechanism attribution and data-quality issues are resolved.

major comments (4)
  1. [§5.7.1, Table 8] The paper attributes a large gain to text-guided correlation filtering: removing the masked distillation branch (MDB) drops DSC by 7.38 points. However, there is no control experiment in which the same branch selects tokens randomly instead of by text-guided correlation. Because the branch is a self-distillation regularizer that shares decoder weights, removes skip connections, masks an overall 12.5% of tokens, and is trained with a Dice consistency loss, the gain could come from the auxiliary decoder, the masking schedule, or the distillation loss itself rather than from the category-specific text prompts. Section 6 explicitly contrasts the method with random-token self-distillation (refs. 53, 54), so this control is necessary to support the paper's central attribution. Please add a random-masking variant with the same K and alpha, and ideally also a variant with a non-informative prompt, and report whether the text-guided selection is indeed load-bearing.
  2. [§4, paragraph 2] The ground-truth masks are mostly generated by SAM2 video propagation rather than manual annotation, with only five manually annotated frames per video. The validation on 10 sampled videos reports an average IoU of 91.1% and manual correction of the 8.3% of propagated masks below 80% IoU, but it is unclear whether that correction was applied only to the 10 sampled videos or to the entire 4,020-image dataset. The residual label noise in the held-out test set is not quantified. If the propagated masks carry systematic bias, all reported DSC/MIoU/OIoU values and model comparisons could be distorted. Please clarify the correction protocol, report the manual/propagated composition of each split, and provide a sensitivity analysis on a fully manually annotated subset.
  3. [§5.5–§5.6, Tables 5–7] The multicentre validation, prospective validation, and gynecologist assessment compare VLM-hyster only against conventional CNN baselines (U-Net, UNet++, Transunet, Swin-UNet, DeepLabV3, PSPNet). They omit the strongest baselines from the main comparison such as Med-SAM, SurgicalSAM, Med-VLM, and Medclip-SAM. As a result, the statements that VLM-hyster 'still obtains the best performance' and demonstrates robustness and generalizability are not supported against the models that actually define the state of the art on the TJ-HS held-out test set. Please add the strongest baselines to these evaluations, or temper the generalization claims accordingly.
  4. [§5.2, Table 2] The paired t-tests and confidence intervals appear to be computed at the level of individual test images, while the TJ-HS split is at the patient level. Images from the same patient are likely correlated, which can inflate statistical significance and produce overly narrow confidence intervals. The paper should state the unit of analysis explicitly and report patient-level bootstrapped confidence intervals or a mixed-effects analysis so that the significance claims are not overstated.
minor comments (5)
  1. [§3.3] The loss equations contain garbled symbols that make the definitions of the distillation loss terms unclear; please restate them with explicit notation for the branch output, backbone output, and ground-truth mask.
  2. [Abstract and §4] The abstract describes 'detailed mask annotations,' which overstates the annotation process since most masks are SAM2-propagated rather than manually drawn. Please revise the wording to distinguish manual and propagated/verified annotations.
  3. [§5.2, Table 2] The comparison protocol for the SAM baseline (DSC 26.73) should be described explicitly, since it is unclear whether SAM is used zero-shot or fine-tuned and what prompts or point inputs were provided.
  4. [§5.7.4, Table 11] The improved masked distillation branch with prompt P3 adds a textual token filtering module; please describe this module's architecture and computational cost in the main text rather than only as a table entry.
  5. [General] There are minor language issues, including 'stablish' in §3.2.2, 'follows a similar manner' in §4, and inconsistent punctuation in the keywords line; these should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity found: the reported gains come from a trained auxiliary branch with fixed text prompts, and no fitted parameter or self-citation is shown to encode the headline result by construction.

full rationale

The paper's central derivation chain is self-contained rather than circular. The masked distillation branch is trained with an explicit loss, L_mbd = Dice(M_mbd, M_seg), where the branch output is produced by a decoder that shares parameters with the segmentation backbone and the text prompts are fixed templates: 'the [category] in hysteroscopic image.' The text-guided correlation filtering selects low-correlation tokens by dot product with the textual [CLS] embedding and masks them at fixed hyperparameters K=50 and alpha=25. These hyperparameters are tuned on the validation set and reported as ablation studies (Section 5.7.1), which is ordinary engineering practice rather than a construction that forces the test-set result. The Table 8 ablation removes the whole branch and shows a drop, which is consistent with the branch contributing to performance, but this is an experimental comparison, not an equation that reduces the prediction to its inputs. The paper does not fit the reported DSC/MIoU/OIoU scores as parameters, nor does it redefine the evaluation metric in terms of the model's own outputs. The reference to MedCLIP is an external pretrained model, and the paper compares against numerous outside baselines; no load-bearing claim relies on a self-citation by the present authors. The reader and skeptic concerns about the missing random-masking control and SAM2-propagated labels are legitimate experimental and data-quality concerns, but they are not circularity: failing to test a random-masking baseline means the text-guidance attribution is not fully proven, not that the derivation is equivalent to its inputs by construction. Therefore the appropriate circularity score is 0.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central result depends on three tuned hyperparameters, the reliability of SAM2-propagated annotations, and the semantic alignment of MedCLIP text and image features. No new physical entities are introduced. These assumptions are stated and partially validated, but they are not independently proven.

free parameters (3)
  • K (percentage of lowest-correlation tokens selected) = 50
    Tuned on the validation set; Section 5.7.1 reports K=50 as optimal.
  • alpha (masking ratio) = 25
    Tuned on the validation set; Section 5.7.1 reports alpha=25 as optimal.
  • lambda (distillation loss weight) = 0.75
    Selected by ablation in Section 5.7.2; lambda=0.75 is the default.
assumptions (4)
  • domain assumption SAM2 video propagation produces masks accurate enough to serve as ground truth for training and evaluation.
    Most annotations are auto-propagated from five manual frames per video. The paper reports 91.1% average IoU on 10 sampled videos and manual correction of 8.3% of masks, but residual label noise is not fully quantified. See Section 4.
  • domain assumption MedCLIP's joint embedding aligns hysteroscopic image patches with category text prompts well enough for the correlation filtering to select meaningful tokens.
    The masked distillation branch depends on this alignment. The paper provides only downstream metrics, not direct analysis of correlation quality. See Section 3.2.
  • domain assumption The text prompt template 'the [category] in hysteroscopic image.' carries sufficient semantic information to guide segmentation.
    Ablations show the prompt wording affects performance, but this establishes sensitivity rather than correctness of the inductive bias. See Section 5.7.4.
  • standard math Dice loss is an appropriate objective for imbalanced multi-category segmentation.
    Widely used in medical segmentation; not controversial. See Section 3.3.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Bootstrapping Vision-Language Model for Hysteroscopic Surgical Scene Segmentation." pith.science (2026). https://pith.science/paper/FUFV4Z3E

@misc{pith2026260809302,
  author       = {Pith},
  title        = {Pith review of: Bootstrapping Vision-Language Model for Hysteroscopic Surgical Scene Segmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FUFV4Z3E}},
  note         = {Machine review of arXiv:2608.09302}
}
read the original abstract

Hysteroscopic surgical scene segmentation plays a pivotal role in understanding the hysteroscopic intraoperative environment as well as computer-assisted intervention. However, this task presents unique challenges due to the high morphological similarity among different lesions and the presence of artifacts such as specular reflections, motion blur, and fluid occlusions in surgical videos. In this work, we propose the first vision-language model (VLM)-based hysteroscopic surgical scene segmentation method, which performs pixel-wise localization for fifteen representative categories in hysteroscopic surgical scenes. Our VLM-hyster has a segmentation backbone that utilizes the pretrained image encoder for robust visual feature extraction, coupled with a transformer-based decoder for dense prediction. Moreover, we design category-specific text prompts and incorporate a masked distillation branch to filter out visual features with low correlation to the text prompts, enabling the model to focus more effectively on category-specific image regions and thereby enhancing segmentation performance. We collect a large multicentric hysteroscopic surgical scene dataset, containing 4,020 high-resolution images with detailed mask annotations, for model training and evaluation. Experimental results demonstrate that VLM-hyster substantially outperforms state-of-the-art AI models. Furthermore, extensive assessments by gynecologists, as well as multicentre and prospective validations, demonstrate VLM-hyster's robustness and generalizability. The results suggest that VLM-hyster earns considerable potential in enabling AI-assisted localization of surgical instruments and lesions in hysteroscopic surgeries. Code is available at https://github.com/viscom-tongji/VLM-hyster.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 7 linked inside Pith

  1. [1]

    YenCF,ChouHH,WuHM,etal.Effectivenessandappropriatenessintheapplicationofoffice hysteroscopy.JournaloftheFormosanMedicalAssociation,2019,118(11):1480-1487

  2. [2]

    Artificial intelligence and surgical decision-making

    Loftus T J, Tighe P J, Filiberto A C, et al. Artificial intelligence and surgical decision-making. JAMAsurgery,2020,155(2):148-158

  3. [3]

    Surgical data science–from concepts toward clinicaltranslation.Medicalimageanalysis,2022,76:102306

    Maier-Hein L, Eisenmann M, Sarikaya D, et al. Surgical data science–from concepts toward clinicaltranslation.Medicalimageanalysis,2022,76:102306

  4. [4]

    AllanM,KondoS,BodenstedtS,etal.2018roboticscenesegmentationchallenge.arXivpreprint arXiv:2001.11190,2020

  5. [5]

    Association of surgical skill assessment with clinical outcomesincancersurgery.JAMAsurgery,2020,155(7):590-598

    Curtis N J, Foster J D, Miskovic D, et al. Association of surgical skill assessment with clinical outcomesincancersurgery.JAMAsurgery,2020,155(7):590-598

  6. [6]

    Towards unified surgical skill assessment

    Liu D, Li Q, Jiang T, et al. Towards unified surgical skill assessment. Proceedings of the IEEE/CVFconferenceoncomputervisionandpatternrecognition,2021:9522-9531

  7. [7]

    Masked-attention mask transformer for universal image segmentation

    Cheng B, Misra I, Schwing A G, et al. Masked-attention mask transformer for universal image segmentation. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition,2022:1290-1299

  8. [8]

    Pseudo-label guided cross-video pixel contrast for robotic surgical scene segmentation with limited annotations

    Yu Y, Zhao Z, Jin Y, et al. Pseudo-label guided cross-video pixel contrast for robotic surgical scene segmentation with limited annotations. 2022 IEEE/RSJ International Conference on IntelligentRobotsandSystems(IROS).IEEE,2022:10857-10864. BiomedicalSignalProcessingandControl 38

Show all 54 references
  1. [9]

    Pixel-wise contrastive learning for multi-class instrument segmentation in endoscopic robotic surgery videos using dataset-wide sample queues[J]

    Sun L, Chen X. Pixel-wise contrastive learning for multi-class instrument segmentation in endoscopic robotic surgery videos using dataset-wide sample queues[J]. IEEE Access, 2024, 12: 156867-156877

  2. [10]

    International Journal of Computer Assisted Radiology andSurgery,2022,17(10):1903-1913

    SunX,ZouY,WangS,etal.Aparallelnetworkutilizinglocalfeaturesandglobalrepresentations for segmentation of surgical instruments. International Journal of Computer Assisted Radiology andSurgery,2022,17(10):1903-1913

  3. [11]

    Deep learning approach for bubble segmentation from hysteroscopicimages.Medical&BiologicalEngineering&Computing,2022,60(6):1613-1626

    Wang D, Dai W, Tang D, et al. Deep learning approach for bubble segmentation from hysteroscopicimages.Medical&BiologicalEngineering&Computing,2022,60(6):1613-1626

  4. [12]

    Radford A,Kim JW,Hallacy C,et al.Learning transferable visual modelsfrom natural language supervision.Internationalconferenceonmachinelearning.PMLR,2021:8748-8763

  5. [13]

    Blip: Bootstrapping language-image pre-training for unified vision- language understanding and generation

    Li J, Li D, Xiong C, et al. Blip: Bootstrapping language-image pre-training for unified vision- language understanding and generation. International conference on machine learning. PMLR, 2022:12888-12900

  6. [14]

    Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing.2024:1081-1093

    XiaP,ZhuK, Li H,et al.Rule: Reliablemultimodal rag forfactualityinmedical visionlanguage models. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing.2024:1081-1093

  7. [15]

    DB-SAM: Delving into High Quality Universal Medical Image Segmentation

    Qin C, Cao J, Fu H, et al. DB-SAM: Delving into High Quality Universal Medical Image Segmentation. International Conference on Medical Image Computing and Computer-Assisted Intervention.Cham:SpringerNatureSwitzerland,2024:498-508

  8. [16]

    Medclip: Contrastive learning from unpaired medical images and text[C]//Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing.2022:3876-3887

    Wang Z, Wu Z, Agarwal D, et al. Medclip: Contrastive learning from unpaired medical images and text[C]//Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing.2022:3876-3887. BiomedicalSignalProcessingandControl 39

  9. [17]

    Unet++: A nested u-net architecture for medicalimagesegmentation.DeepLearninginMedicalImageAnalysisandMultimodalLearning forClinicalDecisionSupport,Springer(2018),pp.3-11

    Zhou Z, Rahman Siddiquee M M, Tajbakhsh N, et al. Unet++: A nested u-net architecture for medicalimagesegmentation.DeepLearninginMedicalImageAnalysisandMultimodalLearning forClinicalDecisionSupport,Springer(2018),pp.3-11

  10. [18]

    Rethinking atrous convolution for semantic image segmentation.arXivpreprintarXiv:1706.05587,2017

    Chen L C, Papandreou G, Schroff F, et al. Rethinking atrous convolution for semantic image segmentation.arXivpreprintarXiv:1706.05587,2017

  11. [19]

    nnU-Net: a self-configuring method for deep learning- basedbiomedicalimagesegmentation[J].Naturemethods,2021,18(2):203-211

    Isensee F, Jaeger P F, Kohl S A A, et al. nnU-Net: a self-configuring method for deep learning- basedbiomedicalimagesegmentation[J].Naturemethods,2021,18(2):203-211

  12. [20]

    Segment anything

    Kirillov A, Mintun E, Ravi N, et al. Segment anything. Proceedings of the IEEE/CVF internationalconferenceoncomputervision.2023:4015-4026

  13. [21]

    Segment anything in medical images[J]

    Ma J, He Y, Li F, et al. Segment anything in medical images[J]. Nature Communications, 2024, 15(1):654

  14. [22]

    Surgicalsam: Efficient class promptable surgical instrument segmentation[C]//Proceedings of the AAAI Conference on Artificial Intelligence

    Yue W, Zhang J, Hu K, et al. Surgicalsam: Efficient class promptable surgical instrument segmentation[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2024, 38(7): 6890-6898

  15. [23]

    SurgVLM: A Large Vision-Language Model and Systematic EvaluationBenchmarkforSurgicalIntelligence[J].arXivpreprintarXiv:2506.02555,2025

    Zeng Z, Zhuo Z, Jia X, et al. SurgVLM: A Large Vision-Language Model and Systematic EvaluationBenchmarkforSurgicalIntelligence[J].arXivpreprintarXiv:2506.02555,2025

  16. [24]

    Med-VLM: Enhancing Medical Image Segmentation Accuracy through Vision-Language Model[C]//Proceedings of the IEEE/CVF International Conference on ComputerVision.2025:7283-7293

    Zhao Y, Zhong E, Yuan C, et al. Med-VLM: Enhancing Medical Image Segmentation Accuracy through Vision-Language Model[C]//Proceedings of the IEEE/CVF International Conference on ComputerVision.2025:7283-7293

  17. [25]

    Automated system for diagnosing endometrial cancer by adoptingdeep-learningtechnologyinhysteroscopy[J].PLoSOne,2021,16(3):e0248526

    Takahashi Y, Sone K, Noda K, et al. Automated system for diagnosing endometrial cancer by adoptingdeep-learningtechnologyinhysteroscopy[J].PLoSOne,2021,16(3):e0248526

  18. [26]

    Automated detection of endometrial polyps from hysteroscopic videosusingdeeplearning[J].Diagnostics,2023,13(8):1409

    Zhao A, Du X, Yuan S, et al. Automated detection of endometrial polyps from hysteroscopic videosusingdeeplearning[J].Diagnostics,2023,13(8):1409. BiomedicalSignalProcessingandControl 40

  19. [27]

    Digital image analysis with fully connected convolutional neural network to facilitate hysteroscopic fibroid resection[J]

    Török P, Harangi B. Digital image analysis with fully connected convolutional neural network to facilitate hysteroscopic fibroid resection[J]. Gynecologic and obstetric investigation, 2018, 83(6): 615-619

  20. [28]

    Exploring intra-and inter-video relation for surgical semantic scene segmentation.IEEETransactionsonMedicalImaging,2022,41(11):2991-3002

    Jin Y, Yu Y, Chen C, et al. Exploring intra-and inter-video relation for surgical semantic scene segmentation.IEEETransactionsonMedicalImaging,2022,41(11):2991-3002

  21. [29]

    Visual instruction tuning

    Liu H, Li C, Wu Q, et al. Visual instruction tuning. Advances in neural information processing systems,2023,36:34892-34916

  22. [30]

    YangA,LiA,YangB,etal.Qwen3technicalreport[J].arXivpreprintarXiv:2505.09388,2025

  23. [31]

    PMLR,2023: 19730-19742

    LiJ,LiD,SavareseS,etal.Blip-2:Bootstrappinglanguage-imagepre-trainingwithfrozenimage encodersand large language models.International conference onmachine learning. PMLR,2023: 19730-19742

  24. [32]

    Text promptable surgical instrument segmentation with vision- languagemodels[J].AdvancesinNeuralInformationProcessingSystems,2023,36:28611-28623

    Zhou Z, Alabi O, Wei M, et al. Text promptable surgical instrument segmentation with vision- languagemodels[J].AdvancesinNeuralInformationProcessingSystems,2023,36:28611-28623

  25. [33]

    Segment anything model for medical image segmentation: Current applicationsandfuturedirections[J].ComputersinBiologyandMedicine,2024,171:108238

    Zhang Y, Shen Z, Jiao R. Segment anything model for medical image segmentation: Current applicationsandfuturedirections[J].ComputersinBiologyandMedicine,2024,171:108238

  26. [34]

    U-net: Convolutional networks for biomedical image segmentation

    Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation. Medical image computing and computer-assisted intervention - MICCAI 2015. Springer,Cham,pp234–241

  27. [35]

    ZhaoR,QianB,ZhangX,etal.Rethinkingdicelossformedicalimagesegmentation.2020IEEE InternationalConferenceonDataMining(ICDM).IEEE,2020:851-860

  28. [36]

    Sam 2: Segment anything in images and videos

    Ravi N, Gabeur V, Hu Y T, et al. Sam 2: Segment anything in images and videos. arXiv preprint arXiv:2408.00714,2024

  29. [37]

    Det-SAM2: Technical Report on the Self-Prompting Segmentation FrameworkBasedonSegmentAnythingModel2.arXivpreprintarXiv:2411.18977,2024

    Wang Z, Zhou Q, Liu Z. Det-SAM2: Technical Report on the Self-Prompting Segmentation FrameworkBasedonSegmentAnythingModel2.arXivpreprintarXiv:2411.18977,2024. BiomedicalSignalProcessingandControl 41

  30. [38]

    Ethical Considerations for Including Children in Clinical Research[J]

    Snyder D. Ethical Considerations for Including Children in Clinical Research[J]. Inside Precision Medicine,2024,11(1):50-53

  31. [39]

    DiceLR.Measuresoftheamountofecologicassociationbetweenspecies.Ecology, 1945,26(3): 297-302

  32. [40]

    Optimizing intersection-over-union in deep neural networks for image segmentation

    Rahman M A, Wang Y. Optimizing intersection-over-union in deep neural networks for image segmentation. International symposium on visual computing. Cham: Springer International Publishing,2016:234-244

  33. [41]

    Transunet: Transformers make strong encoders for medical image segmentation.arXivpreprintarXiv:2102.04306,2021

    Chen J, Lu Y, Yu Q, et al. Transunet: Transformers make strong encoders for medical image segmentation.arXivpreprintarXiv:2102.04306,2021

  34. [42]

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

    Cao H, Wang Y, Chen J, et al. Swin-unet: Unet-like pure transformer for medical image segmentation. European conference on computer vision. Cham: Springer Nature Switzerland, 2022:205-218

  35. [43]

    ZhaoH,ShiJ,QiX,etal.Pyramidsceneparsingnetwork.ProceedingsoftheIEEEconferenceon computervisionandpatternrecognition.2017:2881-2890

  36. [44]

    Image segmentation using text and image prompts[C]//Proceedings of the IEEE/CVFconferenceoncomputervisionandpatternrecognition.2022:7086-7096

    Lüddecke T, Ecker A. Image segmentation using text and image prompts[C]//Proceedings of the IEEE/CVFconferenceoncomputervisionandpatternrecognition.2022:7086-7096

  37. [45]

    ZouX,YangJ,ZhangH,etal.Segmenteverythingeverywhereallatonce[J].Advancesinneural informationprocessingsystems,2023,36:19769-19782

  38. [46]

    Koleilat T, Asgariandehkordi H, Rivaz H, et al. Medclip-sam: Bridging text and image towards universal medical image segmentation[C]//International conference on medical image computing andcomputer-assistedintervention.Cham:SpringerNatureSwitzerland,2024:643-653

  39. [47]

    Journal of Obstetrics andGynaecologyResearch,2014,40(7):1950-1954

    AkdemirA,ErgenogluAM,YenielAÖ,etal.Coring‐typelaparoscopicresectionofacavitated non‐communicating rudimentary horn under hysteroscopic assistance[J]. Journal of Obstetrics andGynaecologyResearch,2014,40(7):1950-1954. BiomedicalSignalProcessingandControl 42

  40. [48]

    Comparative study of the methodologies used for subjective medicalimagequalityassessment.PhysicsinMedicine&Biology,2021,66(15):15TR02

    Lévêque L, Outtas M, Liu H, et al. Comparative study of the methodologies used for subjective medicalimagequalityassessment.PhysicsinMedicine&Biology,2021,66(15):15TR02

  41. [49]

    2025, 39(6):5649-5657

    LiuY,WangG,ZhangJ,et al.Unveilingtheknowledgeofclipfortraining-freeopen-vocabulary semantic segmentation[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2025, 39(6):5649-5657

  42. [50]

    Clearclip: Decomposing clip representations for dense vision- language inference[C]//European Conference on Computer Vision

    Lan M, Chen C, Ke Y, et al. Clearclip: Decomposing clip representations for dense vision- language inference[C]//European Conference on Computer Vision. Cham: Springer Nature Switzerland,2024:143-160

  43. [51]

    Sigmoid loss for language image pre- training[C]//Proceedings of the IEEE/CVF international conference on computer vision

    Zhai X, Mustafa B, Kolesnikov A, et al. Sigmoid loss for language image pre- training[C]//Proceedings of the IEEE/CVF international conference on computer vision. 2023: 11975-11986

  44. [52]

    Pubmedclip: How much does clip benefit visual question answeringinthemedical domain?[C]//FindingsoftheAssociation forComputational Linguistics: EACL2023.2023:1181-1193

    Eslami S, Meinel C, De Melo G. Pubmedclip: How much does clip benefit visual question answeringinthemedical domain?[C]//FindingsoftheAssociation forComputational Linguistics: EACL2023.2023:1181-1193

  45. [53]

    Emerging properties in self-supervised vision transformers[C]//ProceedingsoftheIEEE/CVFinternationalconferenceoncomputervision.2021: 9650-9660

    Caron M, Touvron H, Misra I, et al. Emerging properties in self-supervised vision transformers[C]//ProceedingsoftheIEEE/CVFinternationalconferenceoncomputervision.2021: 9650-9660

  46. [54]

    ZhaoZ,HuangB,XingS,et al.Asymmetricmaskeddistillationfor pre-trainingsmallfoundation models[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2024:18516-18526

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.