Pith. sign in

REVIEW 3 major objections 6 minor 83 references

Recent Advances in Medical Image Classification

T0 review · 3 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 A three-level map of medical image classification: from CNNs and ViTs to Med-VLMs, few-shot/zero-shot learning, and XAI.

desk verdict A broad but uneven survey: the three-level taxonomy is useful, yet the XAI performance claim overreaches and several factual errors need correction. read the letter →

arxiv 2506.04129 v1 pith:PACUPQMS submitted 2025-06-04 eess.IV cs.AIcs.CV

Loan Dao , Ngoc Quoc Ly This is my paper
classification eess.IVcs.AIcs.CV MSC 68T4568T0768T10 PACS 87.57.nm
keywords medicalimageclassificationvision-languagemodelfew-shotlearningzero-shotexplainableAIdeepcomputer-aideddiagnosissurvey
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 is a review paper that tries to organize the field of medical image classification into three nested levels of solutions: basic models (learning paradigms, backbone architectures, and explainability tools), task-specific models (single-task and multitask networks including Med-VLMs), and clinical applications (AI-based CAD across several diseases). The central organizational claim is that these three levels capture the state of the art and that the current frontier is defined by Med-VLMs, few-shot/zero-shot learning, and XAI. A sympathetic reader would take the paper as a structured map of the field, useful for researchers entering the area or looking for a birds-eye view, even though it does not introduce a new scientific result.

What carries the argument

The organizing device is the three-level solution hierarchy (basic models, task-specific models, applications). Within each level, the central technical mechanisms are: vision-language contrastive pretraining (biomedical adaptation of CLIP) that lets Med-VLMs learn from image-caption pairs and enables zero-shot classification; meta-learning and transfer-learning principles for few-shot classification; and gradient-based attribution methods (Integrated Gradients, Grad-CAM, SHAP) from XAI that identify influential features for error correction and model refinement.

What would settle it

Run the exact Integrated-Gradients soft-masking procedure from the two Apicella studies on a standard medical benchmark such as CheXpert or ISIC with a fixed architecture and labeled-data budget; if accuracy does not improve over the baseline without masking, the review's claim that XAI enhances medical classification performance is unsupported.

Watch

Extended reading notes

Core claim

The paper's central claim is organizational: recent advances in medical image classification can be described as progress across three levels of solutions — Level 1 covering fundamental learning models (from supervised to self-supervised and federated learning), Med-VLMs like BiomedCLIP and MedBLIP, fundamental network architectures (CNNs, GNNs, ViTs), and XAI methods for CNNs and transformers; Level 2 covering task-specific architectures for single-task classification and multitask classification-plus-segmentation, including Med-VLMs like GLoRIA and CONCH; and Level 3 covering real-world applications such as breast cancer screening, tuberculosis detection, eye disease diagnosis, skin cancer, and bone disease. The paper argues that the combination of Med-VLMs, few-shot and zero-shot learning, and XAI addresses the field's core obstacle — limited labeled medical data — while also improving model transparency and accuracy.

Load-bearing premise

The conclusion that XAI techniques improve classification accuracy rests on studies using Fashion-MNIST, CIFAR10, and STL10, none of which are medical images; the review generalizes that improvement to medical image classification without direct medical-domain evidence.

Editorial extensions

If this is right

  • A researcher entering medical image classification can use the three-level taxonomy to locate their work — whether they are improving fundamental backbones, designing task-specific multitask models, or building clinical applications.
  • If the surveyed Med-VLM results hold, moving from unimodal CNNs to vision-language pretraining is a viable route when labeled medical data is scarce.
  • If the XAI results transfer to medical domains, then explanation tools such as Integrated Gradients become not just interpretability aids but also training-time mechanisms to improve classification accuracy.
  • Few-shot and zero-shot learning, if adopted clinically, could enable classification of rare or novel diseases that are underrepresented in training datasets.
  • Multitask models that jointly segment and classify may become the default for tasks like tumor analysis, where localization and classification are mutually informative.

Reading between the lines

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

  • The review treats XAI performance gains as if they transfer from natural-image benchmarks to medical imaging, but the cited supporting studies used only Fashion-MNIST, CIFAR10, and STL10; a testable extension would be to run the same Integrated-Gradients masking procedure on chest X-ray or histopathology datasets and measure whether the accuracy gain survives.
  • The three-level taxonomy could be applied as a benchmarking framework: pick one modality (e.g., chest X-ray) and compare a representative model from each level under the same labeled-data budget, which would directly test the review's implicit claim that each level resolves a specific failure mode.
  • By grouping Med-VLMs across both Level 1 and Level 2, the review implies that the same pretraining architecture can serve as a backbone and as a task-specific model; that boundary could be made more explicit and would sharpen the comparison between generic and specialized Med-VLMs.
  • The paper's emphasis on clinical and paraclinical data in Med-VLMs suggests an implicit roadmap: fusing text-based clinical history with imaging data is likely to be a more reliable route to clinically useful classification than image-only approaches.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. This paper is a survey of recent advances in medical image classification (MIC), organized into three solution levels: basic models (learning models, network architectures, XAI), task-specific models (single-task and multi-task classification/segmentation), and applications (AI-CAD systems and specific clinical domains). It covers CNNs, GNNs, Transformers, Med-VLMs, few-shot and zero-shot learning, and XAI, and claims these advances address limited labeled data while improving and explaining predictions. The survey is descriptive rather than technical, summarizing existing literature and providing comparative tables. The central organizational claim is that MIC progress can be mapped through this three-level framework, with recent state-of-the-art directions such as Med-VLMs and XAI as key drivers.

Significance. If the survey's organization and factual summaries are reliable, it offers a useful structured map of a fast-moving field, potentially serving as an entry point for researchers. Strengths include the broad 2020-2024 literature coverage, the explicit three-level taxonomy, comparative tables of models and datasets, and the identification of public medical image databases. However, the survey's reliability is compromised by a load-bearing unsupported claim that XAI improves classification performance, and by several factual errors in descriptions of specific methods (CheXZero, BERTHop) and medical data concepts. These issues need correction before the survey can be recommended as a dependable reference.

major comments (3)
  1. [§III.C.2] The subsection concludes 'XAI explanations enhance both models understanding and classification performance,' but the only cited support for performance improvement is Apicella et al. [67] and [68], both evaluated on Fashion-MNIST, CIFAR10, and STL10, none of which are medical imaging datasets. Since the abstract and conclusion repeat this performance-enhancement claim, the survey generalizes non-medical evidence across a substantial domain shift without support. This claim should either be qualified as a hypothesis requiring medical-domain validation or supported by medical XAI studies that report accuracy gains.
  2. [§III.A.3.b and Table II] CheXZero [9] is listed under 'Zero-shot learning' in Table II, but the text describes it as 'a deep learning model specifically for chest X-ray classification, utilizing pre-trained CNNs and fine-tuning on labelled data.' This contradicts the zero-shot categorization and misrepresents the method: CheXZero is a self-supervised learning approach trained without labels on chest X-rays. The description should be corrected to reflect the actual method.
  3. [Table XII] The BERTHop [38] entry lists 'Zero-shot Learning: AUC: 98.12%' and leaves Few-Shot Learning blank. BERTHop is trained and fine-tuned on labeled chest X-ray data and is not a zero-shot model; reporting a fine-tuned AUC under the zero-shot column conflates evaluation settings and misleads readers about the model's capabilities. The table should separate fine-tuned performance from zero-shot performance or explicitly label the setting.
minor comments (6)
  1. [Table IV] The 'Source' row states 'Direct interaction with healthcare professionals' for both clinical and paraclinical data, but paraclinical data are typically obtained through diagnostic procedures (e.g., imaging, laboratory tests), not direct patient interaction.
  2. [§III.A.2.a] The sentence 'TABLE I. It provides a comparative analysis' appears broken; it likely refers to Table IV rather than Table I, and the 'It' is a fragment.
  3. [§VI.A.1.a] The text references 'Fig. 8' when discussing transfer learning for limited labeled data, but Fig. 8 is the cancer statistics figure from Section V.C; the cross-reference appears misplaced.
  4. [§V.A.2] The database list spells 'ChestX-ray14' as 'ChestX-ray141', and footnote 5 for DeepLesion points to the CAMELYON17 challenge URL rather than the DeepLesion dataset resource.
  5. [Table IX] Reference [29] is cited as 'Grad-CAM for ViTs' but its title describes an XAI-integrated white blood cell identification system; the table entry should be reconciled with the cited work.
  6. [General] There are numerous typos throughout, including 'Explaination' in Table II, 'Outperfrms' and 'iomedical' in Table V, 'Toilored' in Table V, and 'paraclinic al' in the text; a careful proofreading pass is needed.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the survey compiles external literature and makes no derivation that reduces to its own inputs; the lone self-citation is background only.

full rationale

This is a literature survey, not a derivation, so most circularity patterns do not apply. The three-level taxonomy (basic models, task-specific models, applications) is a self-defined organizational scheme, but it is not presented as a predicted or derived result, so no output is equated to an input by construction. The claim in Section III.C.2 that XAI explanations 'enhance both models understanding and classification performance' is supported by two external studies by Apicella et al. [67][68] that evaluate on Fashion-MNIST, CIFAR10, and STL10 rather than medical images; this is an extrapolation or evidence gap, not circularity, because those studies are independent of the authors' framework and are not fitted parameters renamed as predictions. The only self-citation, reference [69], is the authors' prior review of medical image segmentation and is used to support a background statement about progress in segmentation; it does not carry the paper's central claims about Med-VLMs, few-shot learning, zero-shot learning, or XAI. No equations, fitted parameters, or uniqueness theorems are invoked as load-bearing inputs. The paper even cautions in Section VI.A that proposed solutions require further validation in specific clinical contexts, indicating the claims are not being forced by definition. Per the hard rules, an evidence-quality concern about non-medical XAI benchmarks is a correctness risk rather than a circularity finding, so the score remains low.

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

No fitted parameters or invented entities; as a review, the paper's burden is accurate reporting and a valid organizing taxonomy.

assumptions (3)
  • domain assumption Cited papers' reported metrics and model descriptions are accurate as transcribed.
    The survey draws performance numbers and method descriptions from references; any transcription errors propagate (e.g., CheXZero description in Section III.A.3.b).
  • ad hoc to paper The three-level taxonomy (basic, specific, applied) is a meaningful partition of the field.
    The organizing framework is introduced by the authors and not derived from any external standard.
  • domain assumption Findings from non-medical XAI experiments generalize to medical imaging.
    The performance-enhancement claim rests on Apicella et al. studies using Fashion-MNIST, CIFAR10, and STL10 (Section III.C).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Recent Advances in Medical Image Classification." pith.science (2026). https://pith.science/paper/PACUPQMS

@misc{pith2026250604129,
  author       = {Pith},
  title        = {Pith review of: Recent Advances in Medical Image Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PACUPQMS}},
  note         = {Machine review of arXiv:2506.04129}
}
read the original abstract

Medical image classification is crucial for diagnosis and treatment, benefiting significantly from advancements in artificial intelligence. The paper reviews recent progress in the field, focusing on three levels of solutions: basic, specific, and applied. It highlights advances in traditional methods using deep learning models like Convolutional Neural Networks and Vision Transformers, as well as state-of-the-art approaches with Vision Language Models. These models tackle the issue of limited labeled data, and enhance and explain predictive results through Explainable Artificial Intelligence.

Figures

Figures reproduced from arXiv: 2506.04129 by the authors.

Figure 1
Figure 1. Overview of paper organization. TABLE I. LIST OF COMMON ABBREVIATIONS Abbreviation Full Form AI Artificial Intelligence CAD Computer-Aided Diagnosis CNN Convolutional Neural Network CV Computer Vision DL Deep Learning DNN Deep Neural Network FSL Few-shot learning Med-VLM Medical Visual-Language Model MIC Medical Image Classification MTL Multitask Learning ML Machine Learning NLP Natural Language Processing SOTA Stat… view at source ↗
Figure 2
Figure 2. Illustration of convolutional neural networks (CNNs) inspired by biological visual mechanisms [66] [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Illustration of the diverse imaging techniques [70] [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: An overview of the organs and corresponding medical imaging modalities [71] [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: An illustration of the Medical Segmentation decathlon's ten distinct tasks [72]. 2) Public databases in medical imaging research: The growth of public medical image databases has been crucial in advancing disease classification research. Noteworthy databases include: …
Figure 6
Figure 6. Figure 6: Categorization of CAD systems in medical imaging interpretation: a) Second-reader, b) Concurrent-reader, and c) First-reader types [73]. Despite CAD's acknowledged utility, persistent challenges include high development costs, elevated false-positive rates leading to i…
Figure 8
Figure 8. Figure 8: Projected top ten cancer types for new cases and deaths in the United States for 2024, by gender [74]. A. Challenges and Solutions in MIC 1) Medical image data: a) Limited labeled data: Transfer learning has shown promise in addressing the scarcity of labeled data. Kim…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

83 extracted references · 79 canonical work pages

  1. [67]

    Strategies to explo it XAI to improve classification systems,

    A. Apicella, L. Di Lorenzo, F. Isgrò, A. Pollastro, and R. Prevete, “Strategies to explo it XAI to improve classification systems,” in Communications in Computer and Information Science, Cham: Springer Nature Switzerland, 2023, pp. 147–159

    work page 2023

  2. [68]

    An XAI- based masking approach to improve classification systems,

    A. Apicella, S. Giugliano, F. Isgrò, A. Pollastro, and R. Prevete, “An XAI- based masking approach to improve classification systems,” BEWARE@AI*IA, pp. 79–83, 2023

    work page 2023

  3. [69]

    A comprehensive study on medical image segmentation using deep neural networks,

    L. Dao and N. Q. Ly, “A comprehensive study on medical image segmentation using deep neural networks,” Int. J. Adv. Comput. Sci. Appl., vol. 14, no. 3, 2023

    work page 2023

  4. [9]

    Expert-level detection of pathologies from unannotated chest X -ray images via self-supervised learning,

    E. Tiu, E. Talius, P. Patel, C. P. Langlotz, A. Y. Ng, and P. Rajpurkar, “Expert-level detection of pathologies from unannotated chest X -ray images via self-supervised learning,” Nat. Biomed. En g., vol. 6, no. 12, pp. 1399–1406, 2022

    work page 2022

  5. [38]

    BERTHop: An effective vision -and-language model for chest X-ray disease diagnosis,

    M. Monajatipoor, M. Rouhsedaghat, L. H. Li, C. -C. Jay Kuo, A. Chien, and K.-W. Chang, “BERTHop: An effective vision -and-language model for chest X-ray disease diagnosis,” in Lecture Notes in Computer Science, Cham: Springer Nature Switzerland, 2022, pp. 725–734

    work page 2022

  6. [1]

    BiomedCLIP: a multimodal biomedical foundation model pretrained from fifteen million scientific image -text pairs,

    S. Zhang et al., “BiomedCLIP: a multimodal biomedical foundation model pretrained from fifteen million scientific image -text pairs,” arXiv [cs.CV], 2023

    2023

  7. [2]

    XrayGPT: Chest radiographs summarization using medical vision-language models,

    O. Thawkar et al., “XrayGPT: Chest radiographs summarization using medical vision-language models,” arXiv [cs.CV], 2023

    2023

  8. [3]

    M-FLAG: Medical vision-language pre-training with frozen language models and Latent spAce Geometry optimization,

    C. Liu et al., “M-FLAG: Medical vision-language pre-training with frozen language models and Latent spAce Geometry optimization,” arXiv [cs.CV], 2023

    2023

Show all 83 references
  1. [4]

    MedBLIP: Bootstrapping language-image pre-training from 3D medical images and texts,

    Q. Chen, X. Hu, Z. Wang, and Y. Hong, “MedBLIP: Bootstrapping language-image pre-training from 3D medical images and texts,” arXiv [cs.CV], 2023

    2023

  2. [5]

    Lea rning to exploit temporal structure for biomedical vision-language processing,

    S. Bannur et al., “Lea rning to exploit temporal structure for biomedical vision-language processing,” arXiv [cs.CV], 2023

    2023

  3. [6]

    PM2: A new prompting multi-modal model paradigm for few-shot medical image classification,

    Z. Wang, Q. Sun, B. Zhang, P. Wang, J. Zhang, and Q. Zhang, “PM2: A new prompting multi-modal model paradigm for few-shot medical image classification,” arXiv [cs.CV], 2024

    2024

  4. [7]

    DeViDe: Faceted medical knowledge for improved medical vision -language pre- training,

    H. Luo, Z. Zhou, C. Royer, A. Sekuboyina, and B. Menze, “DeViDe: Faceted medical knowledge for improved medical vision -language pre- training,” arXiv [cs.CV], 2024

    2024

  5. [8]

    MedCLIP: Contrastive learning from unpaired medical images and text,

    Z. Wang, Z. Wu, D. Agarwal, and J. Sun, “MedCLIP: Contrastive learning from unpaired medical images and text,” arXiv [cs.CV], 2022

    2022

  6. [10]

    MedKLIP: Medical knowledge enhanced language -image pre-training in radiology,

    C. Wu, X. Zhang, Y. Zhang, Y. Wang, and W. Xie, “MedKLIP: Medical knowledge enhanced language -image pre-training in radiology,” arXiv [eess.IV], 2023

    2023

  7. [11]

    Very deep convolutional networks for large-scale image recognition,

    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” arXiv [cs.CV], 2014

    2014

  8. [12]

    Going Deeper with Convolutions,

    C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, “Going Deeper with Convolutions,” arXiv preprint arXiv:1409.4842, 2014

    2014 arXiv

  9. [13]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. R en, and J. Sun, “Deep residual learning for image recognition,” arXiv [cs.CV], 2015

    2015

  10. [14]

    EfficientNet: Rethinking model scaling for convolutional Neural Networks,

    M. Tan and Q. V. Le, “EfficientNet: Rethinking model scaling for convolutional Neural Networks,” arXiv [cs.LG], 2019

    2019

  11. [15]

    Semi -supervised classification with graph convolutional networks,

    T. N. Kipf and M. Welling, “Semi -supervised classification with graph convolutional networks,” arXiv [cs.LG], 2016

    2016

  12. [16]

    Graph Attention Networks,

    P. Veličković, G. Cucurull, A. Casanova, A. Romero, P. Liò, and Y. Bengio, “Graph Attention Networks,” arXiv [stat.ML], 2017

    2017

  13. [17]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv [cs.CV], 2020

    2020

  14. [18]

    Training data -efficient image transformers & distillation through attention

    Touvron, Hugo, et al. "Training data -efficient image transformers & distillation through attention." International conference on machine learning. PMLR, 2021

    2021

  15. [19]

    TransUNet: Transformers m ake strong encoders for medical image segmentation,

    J. Chen et al., “TransUNet: Transformers m ake strong encoders for medical image segmentation,” arXiv [cs.CV], 2021

    2021

  16. [20]

    TransUNet＋ : Redesigning the skip connection to enhance features in medical image segmentation,

    Y. Liu, H. Wang, Z. Chen, K. Huangliang, and H. Zhang, “TransUNet＋ : Redesigning the skip connection to enhance features in medical image segmentation,” Knowl. Based Syst., vol. 256, no. 109859, p. 109859, 2022

    2022

  17. [21]

    TransU-Net++: Rethinking attention gated TransU-Net for deforestation mapping,

    A. Jamali, S. K. Roy, J. Li, and P. Ghamisi, “TransU-Net++: Rethinking attention gated TransU-Net for deforestation mapping,” Int. J. Appl. Earth Obs. Geoinf., vol. 120, no. 103332, p. 103332, 2023

    2023

  18. [22]

    Why should I trust you?: Explaining the predictions of any classifier,

    M. T. Ribeiro, S. Singh, an d C. Guestrin, “Why should I trust you?: Explaining the predictions of any classifier,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016, pp. 1135–1144

    2016

  19. [23]

    A unified approa ch to interpreting model predictions,

    S. Lundberg and S. -I. Lee, “A unified approa ch to interpreting model predictions,” arXiv [cs.AI], 2017

    2017

  20. [24]

    Learning deep features for discriminative localization,

    B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba, “Learning deep features for discriminative localization,” in 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016

    2016

  21. [25]

    Grad -CAM: Visual explanations from deep networks via Gradient-based localization,

    R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad -CAM: Visual explanations from deep networks via Gradient-based localization,” arXiv [cs.CV], 2016

    2016

  22. [26]

    Grad-CAM++: Improved visual explanations for deep convolutional networks,

    A. Chattopadhyay, A. Sarkar, P. Howlader, and V. N. Balasubramanian, “Grad-CAM++: Improved visual explanations for deep convolutional networks,” arXiv [cs.CV], 2017

    2017

  23. [27]

    ProtoPFormer: Concentrating on prototypical parts in vision transformers for interpretable image recognition,

    M. Xue et al., “ProtoPFormer: Concentrating on prototypical parts in vision transformers for interpretable image recognition,” arXiv [cs.CV], 2022

    2022

  24. [28]

    X -Pruner: eXplainable pruning for vision transformers,

    L. Yu and W. Xiang, “X -Pruner: eXplainable pruning for vision transformers,” arXiv [cs.CV], 2023

    2023

  25. [29]

    An XAI integrated identification system of white blood cell type using variants of vision transformer,

    S. M. Dipto, M. T. Reza, M. N. J. Rahman, M. Z. Parvez, P. D. Barua, and S. Chakraborty, “An XAI integrated identification system of white blood cell type using variants of vision transformer,” in Lecture Notes in Networks and Systems , Cham: Springer Nature Switzerland, 2023,...

    2023

  26. [30]

    U -Net: Convolutional Networks for Biomedical Image Segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U -Net: Convolutional Networks for Biomedical Image Segmentation,” in Lecture Notes in Computer Science, Cham: Springer International Publishing, 2015, pp. 234–241

    2015

  27. [31]

    UNet++: A Nested U-Net Architecture for Medical Image Segmentation,

    Z. Zhou, M. M. Rahman Siddiquee, N. Tajbakhsh, and J. Liang, “UNet++: A Nested U-Net Architecture for Medical Image Segmentation,” in Deep Learning in Medical Image Analysis and Multimod al Learning for Clinical Decision Support, Cham: Springer International Publishing, 2018, pp. 3–11

    2018

  28. [32]

    Malaria Parasite Detection using Efficient Neural Ensembles,

    S. Mishra, “Malaria Parasite Detection using Efficient Neural Ensembles,” j.electron.electromedical.eng.med.inform, vol. 3, no. 3, pp. 119–133, 2021

    2021

  29. [33]

    Image projective transformation rectification with synthetic data for smartphone -captured chest X -ray photos classification,

    C. F. Chong, Y. Wang, B. Ng, W. Luo, and X. Yang, “Image projective transformation rectification with synthetic data for smartphone -captured chest X -ray photos classification,” Comput. Biol. Med., vol. 164, p. 107277, 2023. (IJACSA) International Journal of Advanced Computer...

    2023

  30. [34]

    CCF -GNN: A unified model aggregating appearance, microenvironment, and topology for pathology image classification,

    H. Wang et al., “CCF -GNN: A unified model aggregating appearance, microenvironment, and topology for pathology image classification,” IEEE Trans. Med. Imaging, vol. 42, no. 11, pp. 3179–3193, 2023

    2023

  31. [35]

    GazeGNN: A gaze-guided graph neural network for chest X-ray classification,

    B. Wang et al., “GazeGNN: A gaze-guided graph neural network for chest X-ray classification,” arXiv [cs.CV], 2023

    2023

  32. [36]

    Self -Ensembling Vision Transformer (SEViT) for Robust Medical Image Classification,

    F. Almalik, M. Yaqub, and K. Nandakumar, “Self -Ensembling Vision Transformer (SEViT) for Robust Medical Image Classification,” arXiv [cs.CV], 2022

    2022

  33. [37]

    MedViT: A robust vision transformer for generalized medical image classification,

    O. N. Manzari, H. Ahmadabadi, H. Kashiani, S. B. Shokouhi, and A. Ayatollahi, “MedViT: A robust vision transformer for generalized medical image classification,” Comput. Biol. Med., vol. 157, no. 106791, p. 106791, 2023

    2023

  34. [39]

    Knowledge - enhanced visual-language pre-training on chest radiology images,

    X. Zhang, C. Wu, Y. Zhang, W. Xie, and Y. Wang, “Knowledge - enhanced visual-language pre-training on chest radiology images,” Nat. Commun., vol. 14, no. 1, p. 4542, 2023

    2023

  35. [40]

    CLIPath: Fine-tune CLIP with visual feature fusion for pathology image analysis towards minimizing data collection efforts,

    Z. Lai, Z. Li, L. C. Oliveira, J. Chauhan, B. N. Dugger, and C.-N. Chuah, “CLIPath: Fine-tune CLIP with visual feature fusion for pathology image analysis towards minimizing data collection efforts,” 2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), ...

    2023

  36. [41]

    Contrastive Learning of Medical Visual Representations from Paired Images and Text,

    Y. Zhang, H. Jiang, Y. Miura, C. D. Manning, and C. P. Langlotz, “Contrastive Learning of Medical Visual Representations from Paired Images and Text,” in Proc eedings of the 7th Machine Learning for Healthcare Conference, 05--06 Aug 2022, vol. 182, pp. 2–25

    2022

  37. [42]

    Multitask deep learning for segmentation and classification of primary bone tumors on radiographs,

    C. E. von Schacky et al., “Multitask deep learning for segmentation and classification of primary bone tumors on radiographs,” Radiology, vol. 301, no. 2, pp. 398–406, 2021

    2021

  38. [43]

    One model is all you need: Multi-task learning enables simultaneous histology image segmentation and classification,

    S. Graham et al., “One model is all you need: Multi-task learning enables simultaneous histology image segmentation and classification,” Med. Image Anal., vol. 83, no. 102685, p. 102685, 2023

    2023

  39. [44]

    MNC-Net: Multi- task graph structure learning based on node clustering for early Parkinson’s disease diagnosis,

    L. Huang, X. Ye, M. Yang, L. Pan, and S. H. Zheng, “MNC-Net: Multi- task graph structure learning based on node clustering for early Parkinson’s disease diagnosis,” Comput. Biol. Med., vol. 152, no. 106308, p. 106308, 2023

    2023

  40. [45]

    Attention based multi - task interpretable graph convolutional network for Alzheimer’s disease analysis,

    S. Jiang, Q. Feng, H. Li, Z. Deng, and Q. Jiang, “Attention based multi - task interpretable graph convolutional network for Alzheimer’s disease analysis,” Pattern Recognit. Lett., vol. 180, pp. 1–8, 2024

    2024

  41. [46]

    Transformer -based multi-task learning for classification and segmentation of gastrointestinal tract endoscopic images,

    S. Tang et al., “Transformer -based multi-task learning for classification and segmentation of gastrointestinal tract endoscopic images,” Comput. Biol. Med., vol. 157, no. 106723, p. 106723, 2023

    2023

  42. [47]

    Multi -task approach based on combined CNN-transformer for efficient segmentation and classification of breast tumors in ultrasound ima ges,

    J. Tagnamas, H. Ramadan, A. Yahyaouy, and H. Tairi, “Multi -task approach based on combined CNN-transformer for efficient segmentation and classification of breast tumors in ultrasound ima ges,” Vis. Comput. Ind. Biomed. Art, vol. 7, no. 1, 2024

    2024

  43. [48]

    GLoRIA: A multimodal global -local representation learning framework for label - efficient medical image recognition,

    S.-C. Huang, L. Shen, M. P. Lungren, and S. Yeung, “GLoRIA: A multimodal global -local representation learning framework for label - efficient medical image recognition,” in 2021 IEEE/CVF Internationa l Conference on Computer Vision (ICCV), 2021, pp. 3942–3951

    2021

  44. [49]

    Anatomical Structure-Guided medical vision-language pre- training,

    Q. Li et al., “Anatomical Structure-Guided medical vision-language pre- training,” arXiv [cs.CV], 2024

    2024

  45. [50]

    MeDSLIP: Medical Dual-Stream Language-Image Pre- training for fine-grained alignment,

    F W. Fan et al., “MeDSLIP: Medical Dual-Stream Language-Image Pre- training for fine-grained alignment,” arXiv [cs.CV], 2024

    2024

  46. [51]

    Improving medical vision-language contrastive pretraining with semantics-aware triage,

    B. Liu et al., “Improving medical vision-language contrastive pretraining with semantics-aware triage,” IEEE Trans. Med. Imaging, vol. 42, no. 12, pp. 3579–3589, 2023

    2023

  47. [52]

    A visual -language foundation model for computational pathology,

    M. Y. Lu et al., “A visual -language foundation model for computational pathology,” Nat. Med., vol. 30, no. 3, pp. 863–874, 2024

    2024

  48. [53]

    ECAMP: Entity-centered context-aware Medical Vision language pre-training,

    R. Wang et al., “ECAMP: Entity-centered context-aware Medical Vision language pre-training,” arXiv [cs.CV], 2023

    2023

  49. [54]

    Reduction of false -positive markings on mammograms: A retrospective comparison study using an artificial intelligence -based CAD,

    R. C. Mayo, D. Kent, L. C. Sen, M. Kapoor, J. W. T. Leung, and A. T. Watanabe, “Reduction of false -positive markings on mammograms: A retrospective comparison study using an artificial intelligence -based CAD,” J. Digit. Imaging, vol. 32, no. 4, pp. 618–624, 2019

    2019

  50. [55]

    Artificial intelligence for breast ca ncer screening in mammography (AI -STREAM): Preliminary interim analysis of a prospective multicenter cohort study,

    Y. Chang et al., “Artificial intelligence for breast ca ncer screening in mammography (AI -STREAM): Preliminary interim analysis of a prospective multicenter cohort study,” 2024

    2024

  51. [56]

    Applicability of artificial intelligence -based computer- aided detection (AI –CAD) for pulmonary tuberculosis to comm unity- based active case finding,

    K. Okada et al., “Applicability of artificial intelligence -based computer- aided detection (AI –CAD) for pulmonary tuberculosis to comm unity- based active case finding,” Trop. Med. Health, vol. 52, no. 1, 2024

    2024

  52. [57]

    Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs,

    V. Gulshan et al., “Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs,” JAMA, vol. 316, no. 22, p. 2402, 2016

    2016

  53. [58]

    Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning,

    R. Poplin et al., “Prediction of cardiovascular risk factors from retinal fundus photographs via deep learning,” Nat. Biomed. Eng., vol. 2, no. 3, pp. 158–164, 2018

    2018

  54. [59]

    Dermatologist -level classification of skin cancer with deep neural networks,

    A. Esteva et al., “Dermatologist -level classification of skin cancer with deep neural networks,” Nature, vol. 542, no. 7639, pp. 115–118, 2017

    2017

  55. [60]

    Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists,

    H. A. Haenssle et al., “Man against machine: diagnostic performance of a deep learning convolutional neural network for dermoscopic melanoma recognition in comparison to 58 dermatologists,” Ann. Onc ol., vol. 29, no. 8, pp. 1836–1842, 2018

    2018

  56. [61]

    An intelligent auxiliary framework for bone malignant tumor lesion segmentation in medical image analysis,

    X. Zhan et al., “An intelligent auxiliary framework for bone malignant tumor lesion segmentation in medical image analysis,” Diagnostics (Basel), vol. 13, no. 2, p. 223, 2023

    2023

  57. [62]

    Automated bone tumor segmentation and classification as benign or malignant using computed tomographic imaging,

    I. Yildiz Potter et al., “Automated bone tumor segmentation and classification as benign or malignant using computed tomographic imaging,” J. Digit. Imaging, vol. 36, no. 3, pp. 869–878, 2023

    2023

  58. [63]

    Q. Ye et al., “Automatic detection, segmentation, and classification of primary bone tumors and bon e infections using an ensemble multi -task deep learning framework on multi -parametric MRIs: a multi -center study,” Eur. Radiol., 2023

    2023

  59. [64]

    Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast cancer,

    B. Ehteshami Bejnordi et al., “Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast cancer,” JAMA, vol. 318, no. 22, p. 2199, 2017

    2017

  60. [65]

    Pivotal trial of an autonomous AI -based diagnostic system for detection of diabetic retinopathy in primary care offices,

    M. D. Abràmoff, P. T. Lavin, M. Birch, N. Shah, and J. C. Folk, “Pivotal trial of an autonomous AI -based diagnostic system for detection of diabetic retinopathy in primary care offices,” NPJ Digit. Med., vol. 1, no. 1, p. 39, 2018

    2018

  61. [66]

    Applications of deep learning to neuro -imaging techniques,

    Zhu, G., Jiang, B., Tong, L., Xie, Y., Zaharchuk, G., & Wintermark, M. (2019). Applications of deep learning to neuro -imaging techniques. Frontiers in Neurology, 10, 869. https://doi.org/10.3389/fneur.2019.0086 G. Zhu, B. Jiang, L. Tong, Y. Xie, G. Zaharchuk, and M. Wintermar...

    2019

  62. [70]

    What scans we will read: imaging instrumentation trends in clinical oncology,

    T. Beyer et al., “What scans we will read: imaging instrumentation trends in clinical oncology,” Cancer Imaging, vol. 20, no. 1, pp. 1–38, 2020

    2020

  63. [71]

    LVM-Med: Learning large-scale self-supervised vision models for medical imaging via second -order graph matching,

    D. M. H. Nguyen et al., “LVM-Med: Learning large-scale self-supervised vision models for medical imaging via second -order graph matching,” arXiv [cs.CV], 2023

    2023

  64. [72]

    The Medical Segmentation Decathlon,

    M. Antonelli, A. Reinke, S. Bakas, K. Farahani, and M. Jorge Cardoso, “The Medical Segmentation Decathlon,” Nature Communications, vol. 13, no. 1, p. 4128, 2022

    2022

  65. [73]

    AI -based computer -aided diagnosis (AI -CAD): the latest review to r ead first,

    H. Fujita, “AI -based computer -aided diagnosis (AI -CAD): the latest review to r ead first,” Radiol. Phys. Technol., vol. 13, no. 1, pp. 6 –19, 2020

    2020

  66. [74]

    Cancer statistics, 2024,

    R. L. Siegel, A. N. Giaquinto, and A. Jemal, “Cancer statistics, 2024,” CA Cancer J. Clin., vol. 74, no. 1, pp. 12–49, 2024

    2024

  67. [75]

    Transfer learning for medical image classification: a literature review,

    H. E. Kim, A. Cosa-Linan, N. Santhanam, M. Jannesari, M. E. Maros, and T. Ganslandt, “Transfer learning for medical image classification: a literature review,” BMC Med. Imaging, vol. 22, no. 1, 2022. (IJACSA) International Journal of Advanced Computer Science and Applications,...

    2022

  68. [76]

    CosSIF: Cosine similarity - based image filtering to overcome low inter -class variation in synthetic medical image datasets,

    M. Islam, H. Zunair, and N. Mohammed, “CosSIF: Cosine similarity - based image filtering to overcome low inter -class variation in synthetic medical image datasets,” arXiv [cs.CV], 2023

    2023

  69. [77]

    Semi -supervised learning for medical image classification using imbalanced training data,

    T. Huynh, A. Nibali, and Z. He, “Semi -supervised learning for medical image classification using imbalanced training data,” arXiv [cs.CV], 2021

    2021

  70. [78]

    An effici ent lossless ROI image compression using wavelet-based modified region growing algorithm,

    P. Sreenivasulu and S. Varadarajan, “An effici ent lossless ROI image compression using wavelet-based modified region growing algorithm,” J. Intell. Syst., vol. 29, no. 1, pp. 1063–1078, 2019

    2019

  71. [79]

    Domain adaptation for medical image analysis: A survey,

    H. Guan and M. Liu, “Domain adaptation for medical image analysis: A survey,” arXiv [cs.CV], 2021

    2021

  72. [80]

    Secure, privacy-preserving and federated machine learning in medical imaging,

    G. A. Kaissis, M. R. Makowski, D. Rückert, and R. F. Braren, “Secure, privacy-preserving and federated machine learning in medical imaging,” Nat. Mach. Intell., vol. 2, no. 6, pp. 305–311, 2020

    2020

  73. [81]

    Exploring LRP and Grad- CAM visualization to interpret multi -label-multi-class pathology prediction using chest radiography,

    M. U. Alam, J. R. Baldvinsson, and Y. Wang, “Exploring LRP and Grad- CAM visualization to interpret multi -label-multi-class pathology prediction using chest radiography,” in 2022 IEEE 35th International Symposium on Computer -Based Medical Systems (CBMS), 2022, pp. 258–263

    2022

  74. [82]

    Evalu ation of sampling and cross -validation tuning strategies for regional -scale machine learning classification,

    C. A. Ramezan, T. A. Warner, and A. E. Maxwell, “Evalu ation of sampling and cross -validation tuning strategies for regional -scale machine learning classification,” Remote Sens. (Basel), vol. 11, no. 2, p. 185, 2019

    2019

  75. [83]

    FDA approved Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices: An updated 2022 landscape,

    G. Joshi and M. Bhandari, “FDA approved Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices: An updated 2022 landscape,” Research Square, 2022

    2022

Pith tools

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