Pith. sign in

REVIEW 5 major objections 6 minor 48 references

SD-MAD: Sign-Driven Few-shot Multi-Anomaly Detection in Medical Images

T0 review · 5 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read SD-MAD aligns medical images with LLM-generated radiological sign prompts and an anchor-based inter-anomaly loss to detect and classify multiple anomalies from a few labeled examples, lifting 1-shot subset accuracy from 0.4% to 60.8%.

desk verdict A genuinely new few-shot multi-anomaly task with a reasonable CLIP-based method; evidence is thinner than the claims, and the sign-selection mechanism doesn't consistently help and rests on unproved geometric assumptions. read the letter →

arxiv 2505.16659 v1 pith:FQTCOV7B submitted 2025-05-22 cs.CV

Kaiyu Guo , Tan Pan , Chen Jiang , Zijian Wang , Brian C. Lovell , Limei Han , Yuan Cheng , Mahsa Baktashmotlagh This is my paper · ORCID
classification cs.CV
keywords few-shotanomalydetectionmulti-anomalymedicalimagingvision-languagemodelradiologicalsignspromptselectioninter-anomalyloss
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

The paper tries to establish that few-shot medical anomaly detection can be expanded from a one-class normal-versus-abnormal task to a multi-class task that distinguishes several anomaly types using only a handful of labeled images. Its proposal, SD-MAD, anchors each anomaly category to a set of textual descriptions of radiological signs generated by a large language model, aligns image features to those sign descriptions, and then prunes unreliable signs at inference. The method's most striking evidence is a 1-shot brain MRI experiment where subset accuracy rises from 0.4% for the raw vision-language model to 60.8% for SD-MAD; across seven datasets and three evaluation protocols the paper reports improvements over baselines, with the largest gains in multi-label prediction. If the approach holds up, it would let clinicians separate conditions that need different treatments, such as tumor versus pneumonia, without large annotated datasets.

What carries the argument

The load-bearing object is the normal-prompt anchor $f_{\text{anchor}}$, the text feature of the normal image prompt, which the paper treats as a separating hyperplane between each anomaly category and all others. Training uses a hinge-style loss (Eq. 4) with two terms: it pulls an image's features toward its own category's sign prompts only until they are closer than the anchor, and it pushes other categories' sign prompts away only until they are farther than the anchor. The second mechanism is sign selection at inference: a text feature is kept only if its distance to its own category's decision region is strictly smaller than its distance to every other category (Remark 3.3), and each category's score is the supremum cosine similarity over the surviving signs. The whole pipeline is wrapped in a shift adapter that interpolates learned features with the original transformer features in four image-encoder layers and one text-encoder layer, preserving the vision-language model's prior knowledge while adapting to the few-shot data.

What would settle it

Take a held-out set of abnormal images with known categories, run the trained model, and compute how often the inequality in Remark 3.1 is violated—an image whose distance to its own category's signs exceeds its distance to the normal anchor, or whose anchor distance exceeds its distance to another category. If a substantial share of correctly classified images violates this ordering, the geometric assumption is not the actual source of the accuracy gain; a more direct test would ablate the anchor loss and compare subset accuracy on the same folds.

Watch

Extended reading notes

Core claim

The central claim is that multi-anomaly detection in medical images can be driven by text: each anomaly category is represented by a set of radiological signs written by a large language model, and the model is trained so that an image embedding is closer to its own category's sign embeddings than to a normal-text anchor, which in turn is closer than to any other category's sign embeddings. This anchor-based inter-anomaly loss amplifies the separation between anomaly categories in the shared vision-language embedding space. At inference, a second geometric condition (Remark 3.3) is used to discard sign prompts that lie too far from their own category's decision region, reducing the noise from signs that match multiple categories. The paper argues that together these two mechanisms convert a frozen vision-language model into a few-shot multi-anomaly detector, and reports that on slice 0 of a brain MRI dataset the subset accuracy rises from 0.4% for vanilla CLIP to 60.8% with the full method, while the Hamming score rises from 80.2% to 87.2%.

Load-bearing premise

The method assumes that a single 'normal' text prompt always sits between each anomaly category and every other category in the similarity space: every abnormal image is closer to its own category's signs than to normal, and closer to normal than to any other category's signs.

Editorial extensions

If this is right

  • A single vision-language model can handle multiple anomaly categories at once, so updating the system for a new anomaly type requires only new textual signs, not a new image dataset or retraining from scratch.
  • The sign-selection step makes the detector more robust to imperfect LLM-generated prompts: prompts that match multiple categories or are underfit by few-shot training are discarded automatically at inference.
  • The anchor loss is additive, so it can be combined with pixel-level anomaly detection losses, as the paper does with a pixel-level baseline, to keep segmentation capability while adding multi-category separation.
  • The three proposed protocols—general anomaly detection, multi-label prediction, and category-wise AUROC—define a way to measure multi-anomaly detection performance that existing one-class scoring functions cannot support.

Reading between the lines

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

  • A likely failure mode not tested in the paper: if the normal-prompt anchor does not consistently sit between an anomaly category and the others in real embedding spaces, the anchor loss can pull features in the wrong direction; measuring the rate of Remark 3.1 violations on held-out images would quantify this risk.
  • The sign-selection criterion (Remark 3.3) is defined entirely in text-feature space, so it could be applied to any multimodal classifier that generates multiple textual descriptions per class, such as few-shot object detection from captions, without retraining.
  • The paper's observation that sign selection sometimes hurts category-wise AUROC suggests that outlier prompts can still be informative for some test images; an alternative is to soften selection into a confidence-weighted average rather than a hard discard, which the authors do not explore.
  • The multi-anomaly experiments are built from one anatomical site (brain MRI); extending the same three protocols to chest X-rays or pathology would test whether the sign-alignment mechanism transfers across modalities and sign vocabularies.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper proposes SD-MAD, a few-shot multi-anomaly detection method for medical images built on CLIP. It generates multiple radiological-sign text prompts per anomaly category via an LLM, inserts lightweight shift adapters into selected CLIP encoder layers, and trains with an image-text loss plus an inter-anomaly anchor loss that uses the normal-category text feature as a separating hyperplane. At inference, it selects reliable prompts by re-labeling text features based on nearest-neighbor distances and produces both continuous scores and binary predictions. The authors introduce three evaluation protocols (general AD AUROC, multi-label Hamming/subset accuracy, and category-wise AUROC) and report experiments on seven datasets, with the main multi-anomaly results on brain MRI slices from fastMRI+.

Significance. If validated, the paper makes a useful contribution: it defines a clinically motivated few-shot multi-anomaly task, proposes a concrete VLM adaptation, and evaluates it with three protocols over seven datasets. The use of LLM-generated sign prompts is interesting, and the 1-shot subset-accuracy gains on slice 0 are striking. However, the current evidence is not conclusive because the geometric assumptions underlying the anchor loss and sign selection are unverified and partly contradicted by the paper's own limitation statement, the component contributions are not isolated by ablation, and all few-shot numbers are single runs.

major comments (5)
  1. [§3.2, Remark 3.1 and Eq. (4)] The inter-anomaly loss relies on the stipulated ordering sup_i d(f_c_img, f_c_text,i) ≤ d(f_c_img, f_anchor) ≤ inf_{k≠c,j} d(f_c_img, f^k_text,j). This is asserted rather than derived or verified. If this ordering does not hold for CLIP ViT-L/14 features, the hinge terms in Eq. (4) can push embeddings in directions unrelated to the intended separation, and the binary rule in Eq. (8) has no valid decision boundary. Please provide empirical verification of the ordering on held-out images before and after training (e.g., quantiles of the relevant distance differences), or a proof under the adapter parameterization. This is load-bearing because both training and inference use the anchor.
  2. [§3.3, Remark 3.3 and Eq. (6)] The sign-selection step assumes a labeling function h_text with h_text(f^c_text)=c for every prompt of category c. Yet the Conclusion's Limitation states that 'some prompts may correspond to more than one anomaly type.' For such prompts h_text is not well defined, so Definition 3.2 and Remark 3.3 do not apply to exactly the prompts that sign selection is intended to filter. Please specify how multi-category prompts are assigned, or restrict the selection to prompts for which h_text is a function, and quantify the effect of that restriction on the reported numbers.
  3. [§4.3, Tables 2 and 3] All few-shot results are reported without standard deviations or number of seeds. With K=1, the choice of the single training sample is likely to dominate the outcome. Moreover, the benefit of sign selection is inconsistent: on slice 10 the average category-wise AUROC drops from 67.1% without sign selection to 61.5% with it, and on slice 5 the subset accuracy drops from 29.0% to 27.3%. Please report multiple seeds/random splits with error bars and significance tests, and temper the claim that sign selection yields consistent improvements.
  4. [§4.3, Table 2 and §3.2] The comparison between CLIP/MedCLIP and 'Ours (no SS)' conflates the shift adapter with the inter-anomaly anchor loss. There is no ablation that trains the adapter with only L_img-text, without Lanchor, so the specific contribution of the anchor loss to the observed gains is not established. Please add this ablation, and also report a variant with Lanchor but no adapter if feasible.
  5. [Abstract and Table 1] The abstract claims 'consistent improvements over state-of-the-art baselines' across datasets. Table 1 does not support this at the individual-dataset level: MVFA is higher on BrainMRI (92.4 vs 91.4), RESC (96.2 vs 95.2), and HIS (82.7 vs 81.6), and on all three pixel-level rows MVFA is higher or equal. Only the average is higher. Please either report pairwise significance across runs or revise the claim to 'average improvement.'
minor comments (6)
  1. [§4.2 and Table 1] The baseline naming is inconsistent: the text mentions BRA and BGAD, but the table shows DRA, BGAD, MVFA, and Ours; please clarify which methods are included and use a consistent notation.
  2. [Table 1] The final row has a formatting error: '92.292.5' should be '92.2 / 92.5'.
  3. [§4.3.2] Please explain why the category 'Small vessel chronic white matter ischemic change' cannot be evaluated in slices 5 and 10, as this exclusion affects the comparability of the slice-wise averages.
  4. [Figure 3] The figure lacks axis labels and a caption stating the exact dataset and slice setting; the text mentions the 5th slice but does not restate the protocol.
  5. [Appendix] The paper repeatedly refers to the Appendix for dataset construction, prompt details, and metric definitions, but no appendix is included in the submitted manuscript; please provide it.
  6. [References] References [10]/[11], [23]/[24], and [44]/[45] are duplicated; please consolidate the bibliography.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the training objective, anchor rule, and sign-selection heuristic are internally consistent, but the reported gains are empirical and not reduced to fitted inputs or self-citation.

full rationale

The paper's central claim is that aligning image features with LLM-generated radiological sign prompts, trained on few-shot support samples and evaluated on held-out test images, improves multi-anomaly detection. No step in the derivation chain reduces to its own inputs. The anchor loss (Eq. 4) and binary prediction rule (Eq. 8) both use the same f_anchor, but this is a deliberate training/evaluation consistency, not a fitted parameter renamed as a prediction: the model must generalize to test images, and Table 2 reports held-out accuracy that is not guaranteed by the loss. Remarks 3.1 and 3.3 are stated as geometric postulates, not derived from prior fitted results; the anchor loss attempts to enforce Remark 3.1, and sign selection heuristically filters outlier prompts using only fixed text features. The Limitation admitting that 'some prompts may correspond to more than one anomaly type' is a soundness concern about the labeling function h_text, not a circularity. There are no load-bearing self-citations: references to prior work such as MVFA and MedCLIP are external baselines, and no uniqueness theorem or prior result by the same authors is invoked to force the method's choice. The three evaluation protocols are new metrics/task formulations, not a renaming of a known result. The reported accuracies are empirical outcomes of a trained model on test data, so no 'prediction' is equivalent by construction to its training input.

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

The central method depends on several hand-chosen components: adapter layer placement, interpolation ratio lambda, LLM prompt sets, anchor text choice, and slice selection. These are not fitted to test data, so circularity is low, but they are unexamined degrees of freedom that can materially change results. No new physical entities are proposed.

free parameters (5)
  • Interpolation ratio lambda = 0.1 to 0.9, ablated
    Controls how strongly the shift adapter output influences the CLIP features in Eq. 2; the paper shows robustness to lambda but gives no principled way to choose it.
  • Adapter layer placement = Image layers 6, 8, 18, 24; text last layer
    Chosen by hand to avoid overfitting in the few-shot setting; no systematic search or justification beyond a short statement in Section 4.1.
  • LLM-generated sign prompts = Not listed
    The text descriptions per anomaly category are generated by an unspecified LLM; the exact wording and number of prompts determine the quality of alignment but are not disclosed.
  • Number of prompts per category N_c = Unspecified
    Appears in Eqs. 3 and 4 as a key quantity, but the paper never states how many signs are generated for each category or how they are selected.
  • Anchor text prompt = Text feature for 'normal images'
    The anchor feature f_anchor is fixed as the text prompt corresponding to normal images; this choice drives the binary prediction rule in Eq. 8.
assumptions (5)
  • domain assumption Different anomalies in medical images share common radiological signs within each category and also have unique signs.
    Stated in the abstract and Section 1 as the motivation for using textual sign prompts; the method would lose its foundation if anomaly categories do not have consistent text-describable signs.
  • domain assumption CLIP's large-scale prior knowledge is sufficiently informative for medical images after only light adaptation.
    The method uses CLIP ViT-L/14 with adapters on only four image layers and one text layer, relying on CLIP features transferring to MRI and other medical modalities.
  • ad hoc to paper The normal text prompt feature acts as a separating hyperplane between anomaly categories, as stated in Remark 3.1.
    This geometric property is asserted without proof and directly underpins the anchor loss in Eq. 4 and the binary prediction rule in Eq. 8.
  • ad hoc to paper Outlier sign prompts can be identified by comparing distances between text features, as described in Remark 3.3.
    The inference-time sign selection in Eq. 6 assumes that a prompt belonging to category c is closer to category c than to any other category, and that violated prompts should be discarded.
  • domain assumption The selected fastMRI+ slices 0, 5, and 10 are representative for evaluating multi-anomaly detection.
    Section 4.1 chooses these slices manually and drops one category on slices 5 and 10 because the label is unavailable; the paper does not test robustness to slice choice.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SD-MAD: Sign-Driven Few-shot Multi-Anomaly Detection in Medical Images." pith.science (2026). https://pith.science/paper/FQTCOV7B

@misc{pith2026250516659,
  author       = {Pith},
  title        = {Pith review of: SD-MAD: Sign-Driven Few-shot Multi-Anomaly Detection in Medical Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FQTCOV7B}},
  note         = {Machine review of arXiv:2505.16659}
}
read the original abstract

Medical anomaly detection (AD) is crucial for early clinical intervention, yet it faces challenges due to limited access to high-quality medical imaging data, caused by privacy concerns and data silos. Few-shot learning has emerged as a promising approach to alleviate these limitations by leveraging the large-scale prior knowledge embedded in vision-language models (VLMs). Recent advancements in few-shot medical AD have treated normal and abnormal cases as a one-class classification problem, often overlooking the distinction among multiple anomaly categories. Thus, in this paper, we propose a framework tailored for few-shot medical anomaly detection in the scenario where the identification of multiple anomaly categories is required. To capture the detailed radiological signs of medical anomaly categories, our framework incorporates diverse textual descriptions for each category generated by a Large-Language model, under the assumption that different anomalies in medical images may share common radiological signs in each category. Specifically, we introduce SD-MAD, a two-stage Sign-Driven few-shot Multi-Anomaly Detection framework: (i) Radiological signs are aligned with anomaly categories by amplifying inter-anomaly discrepancy; (ii) Aligned signs are selected further to mitigate the effect of the under-fitting and uncertain-sample issue caused by limited medical data, employing an automatic sign selection strategy at inference. Moreover, we propose three protocols to comprehensively quantify the performance of multi-anomaly detection. Extensive experiments illustrate the effectiveness of our method.

Figures

Figures reproduced from arXiv: 2505.16659 by the authors.

Figure 1
Figure 1. Figures (a) and (b) visualize the difference between our task and previous tasks. Figures (c) [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The pipeline of SD-MAD. In the framework, the training phase is designed to amplify [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The ablation study on λ. We con￾duct the experiments on the multi-label pre￾diction task with two metrics, namely Ham￾ming score and Subset accuracy. To evaluate the ability to recognize the specific anomaly type, we introduce the category-wise AU￾ROC metric. This metric indicates the performance of the model in each category. Since the label Small vessel chronic white matter ischemic change can not be achieved in s… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Visualization of image-text similarity heatmaps. (a) visualizes the heatmap on vanilla CLIP, [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 36 canonical work pages

  1. [1]

    U. Baid, S. Ghodasara, S. Mohan, M. Bilello, E. Calabrese, E. Colak, K. Farahani, J. Kalpathy- Cramer, F. C. Kitamura, S. Pati, et al. The rsna-asnr-miccai brats 2021 benchmark on brain tumor segmentation and radiogenomic classification.arXiv preprint arXiv:2107.02314, 2021

    arXiv 2021

  2. [2]

    Bakas, H

    S. Bakas, H. Akbari, A. Sotiras, M. Bilello, M. Rozycki, J. S. Kirby, J. B. Freymann, K. Farahani, and C. Davatzikos. Advancing the cancer genome atlas glioma mri collections with expert segmentation labels and radiomic features.Scientific data, 4(1):1–13, 2017

    2017

  3. [3]

    J. Bao, H. Sun, H. Deng, Y . He, Z. Zhang, and X. Li. Bmad: Benchmarks for medical anomaly detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4042–4053, 2024

    work page 2024

  4. [4]

    Bassetti, F

    M. Bassetti, F. Magnè, D. R. Giacobbe, L. Bini, and A. Vena. New antibiotics for gram-negative pneumonia.European Respiratory Review, 31(166), 2022

    work page 2022

  5. [5]

    B. E. Bejnordi, M. Veta, P. J. Van Diest, B. Van Ginneken, N. Karssemeijer, G. Litjens, J. A. Van Der Laak, M. Hermsen, Q. F. Manson, M. Balkenhol, et al. Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast cancer.Jama, 318(22):2199–2210, 2017

    work page 2017

  6. [6]

    Bilic, P

    P. Bilic, P. Christ, H. B. Li, E. V orontsov, A. Ben-Cohen, G. Kaissis, A. Szeskin, C. Jacobs, G. E. H. Mamani, G. Chartrand, et al. The liver tumor segmentation benchmark (lits).Medical image analysis, 84:102680, 2023

    work page 2023

  7. [7]

    Y . Cai, H. Chen, X. Yang, Y . Zhou, and K.-T. Cheng. Dual-distribution discrepancy with self-supervised refinement for anomaly detection in medical images.Medical image analysis, 86:102794, 2023

    work page 2023

  8. [8]

    Y . Cao, X. Xu, Y . Cheng, C. Sun, Z. Du, L. Gao, and W. Shen. Personalizing vision-language models with hybrid prompts for zero-shot anomaly detection.IEEE Transactions on Cybernetics, 2025

    work page 2025

Show all 48 references
  1. [9]

    Y . Cao, J. Zhang, L. Frittoli, Y . Cheng, W. Shen, and G. Boracchi. Adaclip: Adapting clip with hybrid learnable prompts for zero-shot anomaly detection. InEuropean Conference on Computer Vision, 2024

    2024

  2. [10]

    C. Ding, G. Pang, and C. Shen. Catching both gray and black swans: Open-set supervised anomaly detection. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 7388–7398, 2022. 10

    2022

  3. [11]

    C. Ding, G. Pang, and C. Shen. Catching both gray and black swans: Open-set supervised anomaly detection. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 7388–7398, 2022

    2022

  4. [13]

    Z. Ding, Q. Dong, H. Xu, C. Li, X. Ding, and Y . Huang. Unsupervised anomaly segmentation for brain lesions using dual semantic-manifold reconstruction. InInternational Conference on Neural Information Processing, pages 133–144. Springer, 2022

    2022

  5. [14]

    Dosovitskiy, L

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

    2010 arXiv

  6. [15]

    Fernando, H

    T. Fernando, H. Gammulle, S. Denman, S. Sridharan, and C. Fookes. Deep learning for medical anomaly detection–a survey.ACM Computing Surveys (CSUR), 54(7):1–37, 2021

    2021

  7. [16]

    M. S. Graham, P.-D. Tudosiu, P. Wright, W. H. L. Pinaya, P. Teikari, A. Patel, J.-M. U-King-Im, Y . H. Mah, J. T. Teo, H. R. Jäger, et al. Latent transformer models for out-of-distribution detection.Medical Image Analysis, 90:102967, 2023

    2023

  8. [17]

    Z. Gu, B. Zhu, G. Zhu, Y . Chen, M. Tang, and J. Wang. Univad: A training-free unified model for few-shot visual anomaly detection.arXiv preprint arXiv:2412.03342, 2024

    2024 arXiv

  9. [18]

    Hassanaly, C

    R. Hassanaly, C. Brianceau, M. Solal, O. Colliot, and N. Burgos. Evaluation of pseudo-healthy image reconstruction for anomaly detection with deep generative models: Application to brain fdg pet.arXiv preprint arXiv:2401.16363, 2024

    2024 arXiv

  10. [19]

    K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770– 778, 2016

    2016

  11. [20]

    J. Hu, Y . Chen, and Z. Yi. Automated segmentation of macular edema in oct using deep neural networks.Medical image analysis, 55:216–227, 2019

    2019

  12. [21]

    L. Hua, X. Su, Y . Luo, S. You, and J. Long. Hieclip: Hierarchical clip with explicit alignment for zero-shot anomaly detection. InICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1–5. IEEE, 2025

    2025

  13. [22]

    Huang, H

    C. Huang, H. Guan, A. Jiang, Y . Zhang, M. Spratling, and Y .-F. Wang. Registration based few-shot anomaly detection. InEuropean conference on computer vision, pages 303–319. Springer, 2022

    2022

  14. [24]

    Huang, A

    C. Huang, A. Jiang, J. Feng, Y . Zhang, X. Wang, and Y . Wang. Adapting visual-language models for generalizable anomaly detection in medical images. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11375–11385, 2024

    2024

  15. [25]

    Jeong, Y

    J. Jeong, Y . Zou, T. Kim, D. Zhang, A. Ravichandran, and O. Dabeer. Winclip: Zero-/few-shot anomaly classification and segmentation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19606–19616, 2023

    2023

  16. [26]

    Jiang, C

    A. Jiang, C. Huang, Q. Cao, S. Wu, Z. Zeng, K. Chen, Y . Zhang, and Y . Wang. Multi-scale cross- restoration framework for electrocardiogram anomaly detection. InInternational Conference on Medical Image Computing and Computer-Assisted Intervention, pages 87–97. Springer, 2023

    2023

  17. [27]

    E. Jin, Q. Feng, Y . Mou, G. Lakemeyer, S. Decker, O. Simons, and J. Stegmaier. Logicad: Explainable anomaly detection via vlm-based text feature extraction. InProceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 4129–4137, 2025. 11

    2025

  18. [28]

    D. S. Kermany, M. Goldbaum, W. Cai, C. C. Valentim, H. Liang, S. L. Baxter, A. McKeown, G. Yang, X. Wu, F. Yan, et al. Identifying medical diagnoses and treatable diseases by image- based deep learning.cell, 172(5):1122–1131, 2018

    2018

  19. [29]

    Landman, Z

    B. Landman, Z. Xu, J. Igelsias, M. Styner, T. Langerak, and A. Klein. Miccai multi-atlas labeling beyond the cranial vault–workshop and challenge. InProc. MICCAI multi-atlas labeling beyond cranial vault—workshop challenge, volume 5, page 12. Munich, Germany, 2015

    2015

  20. [30]

    Linmans, G

    J. Linmans, G. Raya, J. van der Laak, and G. Litjens. Diffusion models for out-of-distribution detection in digital pathology.Medical Image Analysis, 93:103088, 2024

    2024

  21. [31]

    W. Ma, X. Zhang, Q. Yao, F. Tang, C. Wu, Y . Li, R. Yan, Z. Jiang, and S. K. Zhou. Aa-clip: En- hancing zero-shot anomaly detection via anomaly-aware clip.arXiv preprint arXiv:2503.06661, 2025

    2025 arXiv

  22. [32]

    B. H. Menze, A. Jakab, S. Bauer, J. Kalpathy-Cramer, K. Farahani, J. Kirby, Y . Burren, N. Porz, J. Slotboom, R. Wiest, et al. The multimodal brain tumor image segmentation benchmark (brats).IEEE transactions on medical imaging, 34(10):1993–2024, 2014

    1993

  23. [33]

    Montagne, F

    F. Montagne, F. Guisier, N. Venissac, and J.-M. Baste. The role of surgery in lung cancer treatment: present indications and future perspectives—state of the art.Cancers, 13(15):3711, 2021

    2021

  24. [34]

    A. v. d. Oord, Y . Li, and O. Vinyals. Representation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748, 2018

    2018 arXiv

  25. [35]

    Radford, J

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

    2021

  26. [36]

    Schroff, D

    F. Schroff, D. Kalenichenko, and J. Philbin. Facenet: A unified embedding for face recogni- tion and clustering. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 815–823, 2015

    2015

  27. [37]

    Sheynin, S

    S. Sheynin, S. Benaim, and L. Wolf. A hierarchical transformation-discriminating generative model for few shot anomaly detection. InProceedings of the IEEE/CVF international conference on computer vision, pages 8495–8504, 2021

    2021

  28. [38]

    K. Shum, S. Diao, and T. Zhang. Automatic prompt augmentation and selection with chain- of-thought from labeled data. In H. Bouamor, J. Pino, and K. Bali, editors,Findings of the Association for Computational Linguistics: EMNLP 2023, pages 12113–12139, Singapore, Dec

    2023

  29. [39]

    X. Wang, Y . Peng, L. Lu, Z. Lu, M. Bagheri, and R. Summers. Hospital-scale chest x-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. InIEEE CVPR, volume 7, page 46. sn, 2017

    2017

  30. [40]

    Z. Wang, Z. Wu, D. Agarwal, and J. Sun. Medclip: Contrastive learning from unpaired medical images and text. InProceedings of the Conference on Empirical Methods in Natural Language Processing. Conference on Empirical Methods in Natural Language Processing, volume 2022, page 3...

    2022

  31. [41]

    P. Xia, Z. Chen, J. Tian, G. Yangrui, R. Hou, Y . Xu, Z. Wu, Z. Fan, Y . Zhou, K. Zhu, W. Zheng, Z. Wang, X. Wang, X. Zhang, C. Bansal, M. Niethammer, J. Huang, H. Zhu, Y . Li, J. Sun, Z. Ge, G. Li, J. Zou, and H. Yao. CARES: A comprehensive benchmark of trustworthiness in med...

    2024

  32. [42]

    Xiang, Y

    T. Xiang, Y . Zhang, Y . Lu, A. L. Yuille, C. Zhang, W. Cai, and Z. Zhou. Squid: Deep feature in-painting for unsupervised anomaly detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23890–23901, 2023

    2023

  33. [43]

    H. Xu, Y . Zhang, X. Chen, C. Jing, L. Sun, Y . Huang, and X. Ding. Afsc: Adaptive fourier space compression for anomaly detection.IEEE Transactions on Industrial Informatics, 2024. 12

    2024

  34. [45]

    X. Yao, R. Li, J. Zhang, J. Sun, and C. Zhang. Explicit boundary guided semi-push-pull contrastive learning for supervised anomaly detection. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 24490–24499, 2023

    2023

  35. [46]

    Zbontar, F

    J. Zbontar, F. Knoll, A. Sriram, T. Murrell, Z. Huang, M. J. Muckley, A. Defazio, R. Stern, P. Johnson, M. Bruno, M. Parente, K. J. Geras, J. Katsnelson, H. Chandarana, Z. Zhang, M. Drozdzal, A. Romero, M. Rabbat, P. Vincent, N. Yakubova, J. Pinkerton, D. Wang, E. Owens, C. L....

    2018

  36. [47]

    Zhang, Y

    J. Zhang, Y . Xie, G. Pang, Z. Liao, J. Verjans, W. Li, Z. Sun, J. He, Y . Li, C. Shen, et al. Viral pneumonia screening on chest x-rays using confidence-aware anomaly detection.IEEE transactions on medical imaging, 40(3):879–890, 2020

    2020

  37. [48]

    R. Zhao, B. Yaman, Y . Zhang, R. Stewart, A. Dixon, F. Knoll, Z. Huang, Y . W. Lui, M. S. Hansen, and M. P. Lungren. fastmri+, clinical pathology annotations for knee and brain fully sampled magnetic resonance imaging data.Scientific Data, 9(1):152, 2022

    2022

  38. [49]

    K. Zhou, J. Li, W. Luo, Z. Li, J. Yang, H. Fu, J. Cheng, J. Liu, and S. Gao. Proxy-bridged image reconstruction network for anomaly detection in medical images.IEEE Transactions on Medical Imaging, 41(3):582–594, 2021

    2021

  39. [50]

    J. Zhu, C. Ding, Y . Tian, and G. Pang. Anomaly heterogeneity learning for open-set supervised anomaly detection. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 17616–17626, 2024. 13

    2024

  40. [2023]

    Association for Computational Linguistics

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.