Pith. sign in

REVIEW 4 major objections 5 minor 3 cited by

Self-Prompt SAM: Medical Image Segmentation via Automatic Prompt SAM Adaptation

T0 review · 4 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a fully automatic prompt generator—multi-scale masks, boxes, and distance-transform-chosen center points—lets a frozen-weights SAM beat nnUNet on three public 3D medical segmentation benchmarks.

desk verdict A workable SAM adaptation with self-generated prompts, but the 'first' claim is contradicted by the paper's own citations and the AMOS table is mislabeled. read the letter →

arxiv 2502.00630 v1 pith:NYVCRHLB submitted 2025-02-02 cs.CV

Bin Xie , Hao Tang , Dawen Cai , Yan Yan , Gady Agam This is my paper · ORCID
classification cs.CV
keywords medicalimagesegmentationSegmentAnythingModelautomaticpromptgenerationparameter-efficientfine-tuning3Dadaptersmulti-scalegeneratorEuclideandistancetransform
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 tries to show that the Segment Anything Model (SAM), a prompt-driven segmentation foundation model trained on natural images, can be turned into a fully automatic 3D medical image segmenter without any human-supplied prompts. The authors build a self-prompt loop: a multi-scale prompt generator attached to SAM's image encoder produces auxiliary masks, which are then converted into box prompts, mask prompts, and point prompts placed at the mask's most central pixel by a Euclidean distance transform. To make the 2D model work on 3D volumes, they freeze all original SAM weights and insert lightweight depth-fused adapters, depth positional embeddings, and a multi-class adapter. They report that the resulting Self-Prompt-SAM surpasses nnUNet by 2.3% on AMOS2022, 1.6% on ACDC, and 0.5% on Synapse, which is the evidence for the claim that prompt-free SAM adaptation is viable and state of the art.

What carries the argument

The load-bearing mechanism is the self-prompt loop formed by the MSPGenerator and the prompt encoder. The MSPGenerator is a hierarchical convolutional-and-transpose-convolutional decoder that takes five feature maps of different scales from SAM's image encoder and outputs auxiliary multi-class masks with deep supervision. Those masks are turned into three prompt types: one-hot mask prompts, bounding-box prompts, and a single point prompt selected by Euclidean distance transform as the pixel farthest from the mask boundary, i.e. the most central point. The other central block is the DFusedAdapter, an invert-bottleneck adapter with two extra fully connected layers processing the depth dimension and a skip connection, inserted after multi-head self-attention and in parallel with the MLP in every transformer block; it is what lets the frozen 2D SAM read 3D context.

What would settle it

Compute Dice on ACDC with the automatic prompts replaced by ground-truth prompts and, separately, with point prompts selected randomly inside the predicted masks. If the gap between automatic and ground-truth prompts is small, or if the points+boxes+masks combination stops being optimal once masks contain errors, the central claim that the self-prompt design causes the gain collapses.

Watch

Extended reading notes

Core claim

The central discovery claimed is that SAM's own internal features are enough to generate the prompts SAM needs, so the human can be removed from the loop entirely. Starting from multi-scale feature maps of the frozen image encoder, the MSPGenerator predicts auxiliary multi-class masks; those masks are encoded as one-hot semantic prompts, are used to compute bounding-box prompts, and are processed by an Euclidean distance transform to pick the point prompt farthest from any boundary. The authors further claim that among prompt combinations tried on ground truth, points plus boxes plus masks is the most robust, and that the resulting pipeline—with the DFusedAdapter injecting 3D depth information into every transformer block and the MC-Adapter mapping binary outputs to semantic classes—achieves state-of-the-art Dice on AMOS2022, ACDC, and Synapse while keeping every original SAM weight frozen.

Load-bearing premise

The prompt recipe—boxes plus the most central point plus masks—was chosen using perfect ground-truth masks, and the whole method assumes that same recipe stays best when the auxiliary masks contain errors.

Editorial extensions

If this is right

  • If the claims hold, clinicians never need to specify points or boxes: the model generates its own prompts from the image, which is the difference between a research demo and a usable tool.
  • Freezing all original SAM weights means the pre-trained zero-shot knowledge is retained, and only small adapter modules are trained, so adapting to a new medical dataset is parameter-efficient.
  • The reported margins over nnUNet—2.3% on AMOS2022, 1.6% on ACDC, 0.5% on Synapse—imply that prompt-free SAM adaptation can beat dedicated full-training segmentation pipelines on public benchmarks.
  • The same auxiliary masks carry semantic labels through one-hot encoding, giving SAM the multi-class output that binary-mask SAM natively lacks.

Reading between the lines

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

  • A natural test the paper does not run is to feed ground-truth prompts into the same frozen-SAM pipeline and compare with the automatic prompts; the size of that gap would separate prompt-generation quality from SAM's own segmentation ability.
  • The distance-transform 'most central point' rule is a general recipe for point-prompt selection that could transfer to other promptable foundation models beyond SAM, wherever a coarse mask is available.
  • The same self-prompt architecture could be evaluated on non-CT modalities such as MRI and ultrasound, where the MAdapter's modality adaptation and the depth-fused 3D reasoning would face different intensity and anisotropy statistics.
  • If automatic prompts degrade on small or low-contrast organs, a promising extension is to let the mask decoder refine the auxiliary mask and re-generate prompts iteratively, something the current single-pass design does not do.
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 proposes Self-Prompt-SAM, a framework that adapts the pre-trained 2D SAM to 3D medical image segmentation without manual prompts. It introduces a multi-scale prompt generator (MSPGenerator) that outputs auxiliary multi-class masks, from which box prompts, point prompts (selected by Euclidean distance transform), and mask prompts are derived. To handle 3D data, the authors inject a DFusedAdapter into each transformer block of the image encoder and mask decoder, add depth positional embeddings, an input modality adapter (MAdapter), and an MC-Adapter for multi-class output, while keeping the original SAM weights frozen. Experiments are reported on AMOS2022, Synapse, and ACDC with average Dice as the metric, and a nine-stage ablation (S1-S9) is performed on ACDC. The abstract and conclusion claim state-of-the-art results and a 2.3/0.5/1.6 point improvement over nnUNet on the three datasets.

Significance. The automatic-prompt pipeline is practically attractive, and the staged ablation in Table 4 is a genuine strength: it isolates the contribution of the prompt generator, the depth-fused adapter, depth positional embeddings, the modality adapter, and the multi-class adapter. If the experimental claims are verified, the paper would provide a useful recipe for prompt-free SAM deployment in volumetric medical imaging. The reported gains over strong baselines such as nnUNet are plausible in magnitude. However, the current version overstates its novelty and the AMOS evidence is confounded by the Table 1 row label; the central comparisons also lack statistical support and exclude the most relevant auto-prompt baselines.

major comments (4)
  1. [Section 1, Contribution (i), and References [22], [31]] The claim that Self-Prompt-SAM is "the first SAM-based image segmentation framework without any prompts provided" is internally inconsistent with the paper's own reference list. Reference [22] (Auto-prompting SAM for mobile friendly 3D medical image segmentation) and reference [31] (MaskSAM: Towards auto-prompt SAM with mask classification for medical image segmentation, by the same first author) are both SAM-based medical segmentation frameworks that generate prompts automatically and require no manual prompts. Neither method appears in any comparison table, and Appendix B's taxonomy does not place either method in its categories. The novelty claim should be narrowed to the specific mechanism (e.g., joint learning of the prompt generator with the mask decoder, or semantic multi-class output), and the state-of-the-art claim should be tested against these automatic-prompt baselines.
  2. [Section 3.1, Table 1] The AMOS result row is labeled "MaskSAM (Ours)", not "Self-Prompt-SAM (Ours)", while the text and abstract attribute the 2.3% improvement over nnUNet to Self-Prompt-SAM. If the row is the authors' previously published MaskSAM model, then the AMOS claim does not concern the proposed method; if it is a labeling error, it must be corrected. The evaluation protocol also needs clarification: the table caption says "AMOS testing dataset on the leaderboard", whereas Section 3.1 says "all results are based on 5-fold cross-validation without any ensembles".
  3. [Section 2.4, Figure 1] The optimal prompt combination (points + boxes + masks) and the distance-transform center-point rule are selected from experiments in which prompts are generated from ground-truth masks. In the deployed pipeline, prompts come from MSPGenerator auxiliary masks, which are predictions with errors. The paper does not report any experiment that compares prompt-type rankings under automatic prompts, nor does it quantify the gap between ground-truth prompts and automatic prompts. This is load-bearing because the prompt-design choice could change when the auxiliary masks are imperfect. Please add an ablation with MSPGenerator-generated prompts (e.g., points only, boxes only, masks only, and their combinations) and report auxiliary-mask accuracy.
  4. [Tables 2 and 3] The comparisons report a single Dice value per method with no standard deviations, confidence intervals, or significance tests. The claimed Synapse advantage over nnUNet is 86.74 vs. 86.21 on a random 18/12 split, i.e., 0.53 points, which is within the range that could be explained by split variation. The claims of "outperforming" nnUNet and nnFormer are therefore not established at the reported precision. Provide multiple runs or repeated splits with mean±std, or significance tests, for at least the comparisons with margins below 1 point.
minor comments (5)
  1. [Section 3.1, Synapse paragraph] The sentence "our model surpasses SAMed, nnUNet, and nnFormer by 4.9%, 0.5%, 0.0%, and 0.7%" gives four percentages for three methods; according to Table 2, the margins are 4.86, 0.53, and 0.17 points, so the sentence should be corrected.
  2. [Sections 2.1 and 3.2] The adapter name is rendered inconsistently as "DFusedAdapter", "DfusedAdapter", "DFuserAdapter", and "vApdater" (Table 4); please standardize.
  3. [Section 2.4, Figure 1] The text refers to the Myo class "in Figure 1(d)", but Figure 1(d) shows depth incontinuity, not the central-point experiment; the reference should presumably be to Figure 1(b).
  4. [Appendix A] The statement that for empty class channels "we assign the values of both a box prompt and a point prompt to zero" is unclear, since a zero-valued box or point is not a standard SAM prompt; clarify whether the class is skipped or prompted with a placeholder.
  5. [Table 4 and Section 3.2] The ablation is performed only on ACDC; the conclusions about MSPGenerator and DFusedAdapter would be more convincing with at least one additional dataset, especially because the Synapse margins are small.

Circularity Check

1 steps flagged · score 6.0 of 10

AMOS result is the authors' prior MaskSAM result relabeled as Self-Prompt-SAM; the rest of the pipeline is not circular.

  1. renaming known result [Table 1, Section 3.1 (AMOS comparison)]
    "Table 1: "MaskSAM (Ours)" ... 0.901 ... "Table 1 shows that Self-Prompt-SAM outperforms all existing methods in most organs, achieving a new SOTA performance in DSC.""

    The row reporting the AMOS result is labeled 'MaskSAM (Ours)'—the authors' own prior method, reference [31]—while the surrounding text attributes this result to Self-Prompt-SAM. As written, the headline claim of outperforming nnUNet by 2.3% on AMOS2022 is not a measurement made by the Self-Prompt-SAM architecture described in this paper; it is the previously published MaskSAM result re-labeled as the proposed method. The AMOS SOTA 'prediction' therefore reduces to a renamed known result from the authors' earlier work.

full rationale

Apart from the renaming step above, the main derivation chain is not circular. The MSPGenerator is trained with deep supervision against ground-truth labels; the auxiliary masks it produces are used to derive box, point, and mask prompts for the adapted SAM decoder, whose output is compared to ground truth on held-out test sets (AMOS, Synapse, ACDC). No fitted quantity is fed back into the target metric: the final Dice scores are computed on held-out test data with the automatic prompt pipeline, and the prompt-combination choices in Section 2.4 were tuned with oracle GT prompts, not with the test predictions. The Synapse and ACDC results are reported under 'Self-Prompt-SAM (Ours)' and appear to be genuine new measurements with independent content. Separately, and not a circularity issue, the contribution claim of being 'the first SAM-based image segmentation framework without any prompts provided' (Section 1) is contradicted by the paper's own references [22] and [31], including the authors' prior MaskSAM; this is a novelty and integrity concern rather than a circular derivation. The score of 6 reflects that one of the three headline SOTA claims, AMOS2022, is inherited from prior work under a different name.

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

The method rests on standard supervised learning assumptions and on the transferability of frozen SAM weights to medical volumes. Prompt-design choices are validated only with ground-truth prompts, and several architecture hyperparameters are selected by hand without sensitivity analysis.

free parameters (4)
  • Adapter bottleneck ratio (C/4) in DFusedAdapter = C/4
    Chosen by hand in Appendix C, Eq. (C); no ablation studies reported for this ratio.
  • Depth expansion factor (4D) in DFusedAdapter = 4D
    Chosen by hand in Appendix C, Eq. (C); no ablation studies reported.
  • Initial learning rate and poly decay exponent = 0.01, exponent 0.9
    Appendix D; standard medical segmentation settings, but no sensitivity analysis provided.
  • Warm-up length for MSPGenerator-only training = 200 epochs
    Appendix D sets the first 200 epochs to train only the deep supervision of MSPGenerator; no ablation on this schedule length.
assumptions (5)
  • domain assumption Pre-trained SAM weights retain useful representations for medical imaging after lightweight adapters are added.
    Section 2 keeps all SAM weights frozen and only adds blocks; the whole method assumes transfer works.
  • domain assumption 3D spatial continuity can be recovered by depth-wise adapters and depth positional embeddings.
    Sections 2.1-2.2; DFusedAdapter and DPosEmbed are asserted to fix depth incontinuity shown in Fig. 1(d), with only ablation evidence on ACDC.
  • domain assumption The Euclidean distance transform picks the best point prompt.
    Section 2.4 and Fig. 1(b) support this empirically on an unspecified set of images, not with a formal argument.
  • ad hoc to paper Prompt design chosen under ground-truth prompts remains best for automatic prompts.
    Section 2.4 selects points + boxes + masks using ground-truth prompts; no test compares this against automatic prompts.
  • domain assumption Benchmark ground-truth annotations are correct and preprocessing is equivalent across compared methods.
    Section 3 compares Dice scores across methods; comparability depends on this assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Self-Prompt SAM: Medical Image Segmentation via Automatic Prompt SAM Adaptation." pith.science (2026). https://pith.science/paper/NYVCRHLB

@misc{pith2026250200630,
  author       = {Pith},
  title        = {Pith review of: Self-Prompt SAM: Medical Image Segmentation via Automatic Prompt SAM Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NYVCRHLB}},
  note         = {Machine review of arXiv:2502.00630}
}
read the original abstract

Segment Anything Model (SAM) has demonstrated impressive zero-shot performance and brought a range of unexplored capabilities to natural image segmentation tasks. However, as a very important branch of image segmentation, the performance of SAM remains uncertain when applied to medical image segmentation due to the significant differences between natural images and medical images. Meanwhile, it is harsh to meet the SAM's requirements of extra prompts provided, such as points or boxes to specify medical regions. In this paper, we propose a novel self-prompt SAM adaptation framework for medical image segmentation, named Self-Prompt-SAM. We design a multi-scale prompt generator combined with the image encoder in SAM to generate auxiliary masks. Then, we use the auxiliary masks to generate bounding boxes as box prompts and use Distance Transform to select the most central points as point prompts. Meanwhile, we design a 3D depth-fused adapter (DfusedAdapter) and inject the DFusedAdapter into each transformer in the image encoder and mask decoder to enable pre-trained 2D SAM models to extract 3D information and adapt to 3D medical images. Extensive experiments demonstrate that our method achieves state-of-the-art performance and outperforms nnUNet by 2.3% on AMOS2022, 1.6% on ACDCand 0.5% on Synapse datasets.

Figures

Figures reproduced from arXiv: 2502.00630 by the authors.

Figure 1
Figure 1. (a) Predictions of different combinations of points, boxes, and masks from ground [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The overview architecture of the proposed Self-Prompt-SAM. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison on the Synapse and ACDC dataset. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Hierarchical Self-Prompting SAM: A Prompt-Free Medical Image Segmentation Framework

    cs.CV 2025-06 conditional novelty 5.0 of 10

    HSP-SAM adds learned abstract prompt pairs to SAM, achieving prompt-free medical image segmentation with reported zero-shot improvements of up to 14.04 percent Dice.

  2. Prompt Mechanisms in Medical Imaging: A Comprehensive Survey

    eess.IV 2025-06 conditional novelty 4.0 of 10

    A broad survey that organizes prompt mechanisms for medical image generation, segmentation, and classification into a two-dimensional taxonomy of core technologies and clinical applications.

  3. Prompt Engineering in Segment Anything Model: Methodologies, Applications, and Emerging Challenges

    cs.CV 2025-07 conditional novelty 2.0 of 10

    A structured survey of prompt engineering methods for the Segment Anything Model, covering geometric, textual, and multimodal prompts and their applications.

Reference graph

Works this paper leans on

36 extracted references · 14 canonical work pages · cited by 3 Pith papers

  1. [22]

    arXiv preprint arXiv:2308.14936 (2023) 3

    Li, C., Khanduri, P., Qiang, Y., Sultan, R.I., Chetty, I., Zhu, D.: Auto- prompting sam for mobile friendly 3d medical image segmentation. arXiv preprint arXiv:2308.14936 (2023) 3

    arXiv 2023

  2. [31]

    arXiv preprint arXiv:2403.14103 (2024) 3

    Xie, B., Tang, H., Duan, B., Cai, D., Yan, Y.: Masksam: Towards auto-prompt sam with mask classification for medical image segmentation. arXiv preprint arXiv:2403.14103 (2024) 3

    arXiv 2024

  3. [1]

    Journal of digital imaging 30(4), 449–459 (2017) 1

    Akkus, Z., Galimzianova, A., Hoogi, A., Rubin, D.L., Erickson, B.J.: Deep learning for brain mri segmentation: state of the art and future directions. Journal of digital imaging 30(4), 449–459 (2017) 1

    work page 2017

  4. [2]

    Bernard, O., Lalande, A., Zotti, C., Cervenansky, F., Yang, X., Heng, P.A., Cetin, I., Lekadir, K., Camara, O., Ballester, M.A.G., et al.: Deep learning techniques for automatic mri cardiac multi-structures segmentation and diagnosis: Is the problem solved? IEEE TMI (2018) 1, 4, 6

    work page 2018

  5. [3]

    NeurIPS (2020) 1

    Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J.D., Dhariwal, P., Nee- lakantan, A., Shyam, P., Sastry, G., Askell, A., et al.: Language models are few-shot learners. NeurIPS (2020) 1

    work page 2020

  6. [4]

    arXiv preprint arXiv:2309.03493 (2023) 6, 7

    Bui, N.T., Hoang, D.H., Tran, M.T., Le, N.: Sam3d: Segment anything model in volumetric medical images. arXiv preprint arXiv:2309.03493 (2023) 6, 7

    arXiv 2023

  7. [5]

    arXiv preprint arXiv:2105.05537 (2021) 6, 7

    Cao, H., Wang, Y., Chen, J., Jiang, D., Zhang, X., Tian, Q., Wang, M.: Swin- unet: Unet-like pure transformer for medical image segmentation. arXiv preprint arXiv:2105.05537 (2021) 6, 7

    arXiv 2021

  8. [6]

    arXiv preprint arXiv:2102.04306 (2021) 1, 6, 7

    Chen, J., Lu, Y., Yu, Q., Luo, X., Adeli, E., Wang, Y., Lu, L., Yuille, A.L., Zhou, Y.: Transunet: Transformers make strong encoders for medical image segmentation. arXiv preprint arXiv:2102.04306 (2021) 1, 6, 7

    arXiv 2021

Show all 36 references
  1. [7]

    arXiv preprint arXiv:2304.04155 (2023) 2

    Deng, R., Cui, C., Liu, Q., Yao, T., Remedios, L.W., Bao, S., Landman, B.A., Wheless, L.E., Coburn, L.A., Wilson, K.T., et al.: Segment anything model (sam) for digital pathology: Assess zero-shot segmentation on whole slide imaging. arXiv preprint arXiv:2304.04155 (2023) 2

    2023 arXiv

  2. [8]

    arXiv preprint arXiv:2010.11929 (2020) 2, 7

    Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al.: An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929 (2020) 2, 7

    2020 arXiv

  3. [9]

    arXiv preprint arXiv:2306.13465 (2023) 3

    Gong, S., Zhong, Y., Ma, W., Li, J., Wang, Z., Zhang, J., Heng, P.A., Dou, Q.: 3dsam-adapter: Holistic adaptation of sam from 2d to 3d for promptable medical image segmentation. arXiv preprint arXiv:2306.13465 (2023) 3

    2023 arXiv

  4. [10]

    In: International MICCAI Brainlesion Workshop (2021) 6, 7

    Hatamizadeh, A., Nath, V., Tang, Y., Yang, D., Roth, H.R., Xu, D.: Swin unetr: Swin transformers for semantic segmentation of brain tumors in mri images. In: International MICCAI Brainlesion Workshop (2021) 6, 7

    2021

  5. [11]

    In: W ACV (2022) 6, 7

    Hatamizadeh, A., Tang, Y., Nath, V., Yang, D., Myronenko, A., Landman, B., Roth, H.R., Xu, D.: Unetr: Transformers for 3d medical image segmentation. In: W ACV (2022) 6, 7

    2022

  6. [12]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    He, K., Chen, X., Xie, S., Li, Y., Doll´ ar, P., Girshick, R.: Masked autoencoders are scalable vision learners. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 16000–16009 (2022) 2

    2022

  7. [13]

    In: International Conference on Machine Learning

    Houlsby, N., Giurgiu, A., Jastrzebski, S., Morrone, B., De Laroussilhe, Q., Ges- mundo, A., Attariyan, M., Gelly, S.: Parameter-efficient transfer learning for nlp. In: International Conference on Machine Learning. pp. 2790–2799. PMLR (2019) 2, 3, 8

    2019

  8. [14]

    arXiv preprint arXiv:2304.08506 (2023) 2

    Hu, C., Li, X.: When sam meets medical images: An investigation of segment anything model (sam) on multi-phase liver tumor segmentation. arXiv preprint arXiv:2304.08506 (2023) 2

    2023 arXiv

  9. [15]

    arXiv preprint arXiv:2106.09685 (2021) 2, 3 Self-Prompt SAM 11

    Hu, E.J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., Wang, L., Chen, W.: Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685 (2021) 2, 3 Self-Prompt SAM 11

    2021 arXiv

  10. [16]

    arXiv preprint arXiv:1904.08128 (2019) 1, 6, 7

    Isensee, F., J¨ ager, P.F., Kohl, S.A., Petersen, J., Maier-Hein, K.H.: Automated design of deep learning methods for biomedical image segmentation. arXiv preprint arXiv:1904.08128 (2019) 1, 6, 7

    2019 arXiv

  11. [17]

    arXiv preprint arXiv:2304.06022 (2023) 2

    Ji, G.P., Fan, D.P., Xu, P., Cheng, M.M., Zhou, B., Van Gool, L.: Sam strug- gles in concealed scenes–empirical study on” segment anything”. arXiv preprint arXiv:2304.06022 (2023) 2

    2023 arXiv

  12. [18]

    arXiv preprint arXiv:2304.05750 (2023) 2

    Ji, W., Li, J., Bi, Q., Li, W., Cheng, L.: Segment anything is not always per- fect: An investigation of sam on different real-world applications. arXiv preprint arXiv:2304.05750 (2023) 2

    2023 arXiv

  13. [19]

    NeurIPS (2022) 1, 4, 6

    Ji, Y., Bai, H., Ge, C., Yang, J., Zhu, Y., Zhang, R., Li, Z., Zhanng, L., Ma, W., Wan, X., et al.: Amos: A large-scale abdominal multi-organ benchmark for versatile medical image segmentation. NeurIPS (2022) 1, 4, 6

    2022

  14. [20]

    arXiv preprint arXiv:2304.02643 (2023) 2

    Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A.C., Lo, W.Y., et al.: Segment anything. arXiv preprint arXiv:2304.02643 (2023) 2

    2023 arXiv

  15. [21]

    In: Proc

    Landman, B., Xu, Z., Igelsias, J.E., Styner, M., Langerak, T., Klein, A.: Miccai multi-atlas labeling beyond the cranial vault–workshop and challenge. In: Proc. MICCAI: Multi-Atlas Labeling Beyond Cranial Vault-Workshop Challenge (2015) 1, 4, 6

    2015

  16. [23]

    arXiv preprint arXiv:2304.12306 (2023) 3

    Ma, J., Wang, B.: Segment anything in medical images. arXiv preprint arXiv:2304.12306 (2023) 3

    2023 arXiv

  17. [24]

    arXiv preprint arXiv:2304.04738 2, 4 (2023) 2

    Mohapatra, S., Gosai, A., Schlaug, G.: Sam vs bet: A comparative study for brain extraction and segmentation of magnetic resonance images using deep learning. arXiv preprint arXiv:2304.04738 2, 4 (2023) 2

    2023 arXiv

  18. [25]

    OpenAI: GPT-4 technical report (2023) 1

    2023

  19. [26]

    NeurIPS (2022) 2

    Pan, J., Lin, Z., Zhu, X., Shao, J., Li, H.: St-adapter: Parameter-efficient image- to-video transfer learning. NeurIPS (2022) 2

    2022

  20. [27]

    In: International Conference on Medical image computing and computer-assisted intervention (2015) 1, 7

    Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomedi- cal image segmentation. In: International Conference on Medical image computing and computer-assisted intervention (2015) 1, 7

    2015

  21. [28]

    In: MICCAI (2021) 6

    Wang, W., Chen, C., Ding, M., Yu, H., Zha, S., Li, J.: Transbts: Multimodal brain tumor segmentation using transformer. In: MICCAI (2021) 6

    2021

  22. [29]

    arXiv preprint arXiv:2304.12620 (2023) 3

    Wu, J., Fu, R., Fang, H., Liu, Y., Wang, Z., Xu, Y., Jin, Y., Arbel, T.: Medical sam adapter: Adapting segment anything model for medical image segmentation. arXiv preprint arXiv:2304.12620 (2023) 3

    2023 arXiv

  23. [30]

    In: Proceedings of the Asian Conference on Com- puter Vision

    Xie, B., Tang, H., Cai, D., Yan, Y.: Ms-umlp: Medical image segmentation via multi-scale u-shape mlp-mixer. In: Proceedings of the Asian Conference on Com- puter Vision. pp. 1793–1808 (2024) 1

    2024

  24. [32]

    arXiv preprint arXiv:2107.08623 (2021) 7

    Xu, G., Wu, X., Zhang, X., He, X.: Levit-unet: Make faster encoders with trans- former for medical image segmentation. arXiv preprint arXiv:2107.08623 (2021) 7

    2021 arXiv

  25. [33]

    arXiv preprint arXiv:2302.03024 (2023) 2, 3, 4 12 B

    Yang, T., Zhu, Y., Xie, Y., Zhang, A., Chen, C., Li, M.: Aim: Adapting image mod- els for efficient video action recognition. arXiv preprint arXiv:2302.03024 (2023) 2, 3, 4 12 B. Xie et al

    2023 arXiv

  26. [34]

    arXiv preprint arXiv:2304.13785 (2023) 6, 7

    Zhang, K., Liu, D.: Customized segment anything model for medical image seg- mentation. arXiv preprint arXiv:2304.13785 (2023) 6, 7

    2023 arXiv

  27. [35]

    arXiv preprint arXiv:2109.03201 (2021) 1, 6, 7

    Zhou, H.Y., Guo, J., Zhang, Y., Yu, L., Wang, L., Yu, Y.: nnformer: Interleaved transformer for volumetric segmentation. arXiv preprint arXiv:2109.03201 (2021) 1, 6, 7

    2021 arXiv

  28. [36]

    Zhou, T., Zhang, Y., Zhou, Y., Wu, Y., Gong, C.: Can sam segment polyps? arXiv preprint arXiv:2304.07583 (2023) 2 Self-Prompt SAM 1 A Rethinking SAM. SAM is the first prompt-driven foundation model for natural image segmenta- tion, which is trained on the large-scale SA-1B dat...

    2023 arXiv

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.