pith. sign in

arxiv: 2605.05034 · v1 · submitted 2026-05-06 · 💻 cs.CV

Few-Shot Learning Pipeline for Monkeypox Skin Disease Classification Using CNN Feature Extractors

Md. Safirur Rashid , Sabbir Ahmed , Muhammad Usama Islam , Sumona Hoque Mumu , Md. Hasanul Kabir This is my paper

Pith reviewed 2026-05-08 17:41 UTC · model grok-4.3

classification 💻 cs.CV
keywords few-shot learningmonkeypox classificationCNN feature extractorsSimpleShotskin disease recognitioninductive learningdomain shiftmedical imaging
0
0 comments X

The pith

A simple few-shot classifier using frozen CNN features accurately identifies monkeypox skin lesions from limited examples.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that few-shot learning can overcome the data scarcity problem for classifying rare or emerging diseases like monkeypox by passing images through frozen pretrained CNN backbones to obtain embeddings and then applying SimpleShot nearest-centroid classification in normalized space. A sympathetic reader would care because traditional CNNs require large annotated datasets that are impractical to collect for new conditions. The work benchmarks six CNN backbones under consistent 2-way, 4-way, and 6-way tasks with 1-shot, 5-shot, and 10-shot settings on three public datasets and finds that MobileNetV2_100 consistently leads. Cross-dataset tests reveal stable performance for binary Mpox-versus-others tasks but sharp drops in multi-class settings under domain shift. These outcomes point to the practical value of lightweight inductive methods paired with domain-aware design for clinical tools.

Core claim

The central claim is that a few-shot learning pipeline which extracts features from skin lesion images using frozen pretrained CNN backbones and classifies them with SimpleShot via nearest-centroid comparisons in a normalized embedding space achieves strong performance on monkeypox and pox-like disease recognition tasks. Among six backbones tested, MobileNetV2_100 consistently yields the highest accuracy in 2-way, 4-way, and 6-way settings with 1, 5, and 10 shots on three public datasets. Cross-dataset evaluation further shows stable binary Mpox-vs-Others transfer but significant degradation in multi-class performance under domain shift.

What carries the argument

SimpleShot, the non-parametric inductive classifier that assigns labels based on nearest-centroid distances in the normalized feature space produced by frozen CNN backbones.

If this is right

  • MobileNetV2_100 generates the most effective features for this few-shot skin-lesion task among the six backbones examined.
  • Binary monkeypox-versus-others classification transfers more stably across datasets than multi-class setups.
  • Inductive few-shot methods combined with lightweight CNN backbones offer a workable route for medical classification when labeled examples are scarce.
  • Domain shift affects multi-class performance far more than binary performance, so robustness checks are required for deployment.
  • The overall pipeline supports reliable real-world clinical use once domain variability is taken into account.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same frozen-backbone plus SimpleShot pattern could be tested on other data-scarce dermatological conditions such as other viral exanthems.
  • Adding lightweight domain-adaptation steps might reduce the multi-class degradation observed under dataset shift.
  • Real hospital images likely contain greater lighting, skin-tone, and imaging-device variability than the public collections, providing a direct test of the embedding-discriminability assumption.
  • The emphasis on domain robustness suggests that future pipelines should include explicit cross-site or cross-camera validation as a standard step.

Load-bearing premise

That feature embeddings from frozen ImageNet-pretrained CNN backbones remain sufficiently discriminative for pox-like skin lesions under few-shot regimes and that the three public datasets adequately represent real clinical variability.

What would settle it

A new clinical dataset of monkeypox and similar lesions on which the same frozen-backbone plus SimpleShot pipeline yields substantially lower accuracy than reported on the public sets, or where MobileNetV2_100 no longer ranks first.

Figures

Figures reproduced from arXiv: 2605.05034 by Md. Hasanul Kabir, Md. Safirur Rashid, Muhammad Usama Islam, Sabbir Ahmed, Sumona Hoque Mumu.

Figure 1
Figure 1. Figure 1: Overview of the proposed pipeline: input images are processed by a frozen CNN view at source ↗
Figure 2
Figure 2. Figure 2: Example images from the MSLD v1.0 dataset [ view at source ↗
Figure 3
Figure 3. Figure 3: Example images from the MSLD v2.0 [9] dataset. (a) Monkeypox: well￾circumscribed pustules with central umbilication. (b) Chickenpox: superficial vesicles on red bases at mixed stages. (c) Measles: widespread confluent maculopapular rash. (d) Cowpox: necrotic lesion with central eschar. (e) HFMD: vesicles localized to palms, soles, or mucosa. (f) Healthy: normal skin without lesions, serving as control. Rep… view at source ↗
Figure 4
Figure 4. Figure 4: Samples from the MSID [14] dataset. (a) Monkeypox: firm pustules with umbil￾ication and surrounding erythema. (b) Chickenpox: fragile vesicles at different stages. (c) Measles: diffuse erythematous maculopapular rash. (d) Healthy: skin without visible lesions or inflammation. Although MSID was later expanded to over 8,600 samples through aggressive aug￾mentation, our experiments rely solely on the original… view at source ↗
Figure 5
Figure 5. Figure 5: Accuracy improvement with increasing support size on the 6-way classification view at source ↗
Figure 6
Figure 6. Figure 6: Class-wise accuracy using the MobileNetV2_100 backbone across 2-way, 4-way, and 6-way tasks. Accuracy for Monkeypox decreases as the number of classes increases, indicating greater inter-class confusion in more complex settings view at source ↗
Figure 7
Figure 7. Figure 7: Confusion matrix on the 6-way MSLD v2.0 task (10-shot). Misclassifications view at source ↗
read the original abstract

Despite the strong performance of Convolutional Neural Networks (CNNs) in disease classification, their effectiveness often depends on access to large annotated datasets, which is an impractical requirement for emerging or rare conditions such as Monkeypox. To overcome this limitation, we propose a few-shot learning (FSL) framework that employs SimpleShot, a lightweight, non-parametric, inductive classifier, for Monkeypox and pox-like skin disease recognition from limited labeled examples. The proposed pipeline passes the skin lesion images through a frozen, pretrained CNN backbone to obtain feature embeddings, which are then classified via SimpleShot using nearest-centroid comparisons in a normalized embedding space. We systematically benchmark six widely used CNN backbones as feature extractors under consistent experimental settings, enabling fair comparison. Experiments on three publicly available datasets (MSLD v1.0, MSID, and MSLD v2.0) are conducted across 2-way, 4-way, and 6-way tasks with 1-shot, 5-shot, and 10-shot configurations. Among all models, MobileNetV2_100 consistently achieves the highest accuracy. In addition, we present a cross-dataset evaluation for Monkeypox classification, revealing that binary Mpox-vs-Others transfer remains comparatively stable while multi-class performance degrades significantly under domain shift. Together, these results demonstrate the practical utility of combining inductive FSL methods with lightweight CNN backbones and highlight the importance of domain robustness for reliable real-world clinical deployment.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper proposes a few-shot learning pipeline for classifying Monkeypox and similar skin diseases. It extracts features from lesion images using frozen ImageNet-pretrained CNN backbones and classifies them with the non-parametric SimpleShot method via nearest-centroid matching in normalized space. Six backbones are benchmarked under a consistent protocol on three public datasets (MSLD v1.0, MSID, MSLD v2.0) for 2/4/6-way tasks at 1/5/10 shots. MobileNetV2_100 is reported to achieve the highest accuracy. Cross-dataset transfer experiments show relatively stable binary Mpox-vs-Others performance but marked degradation in multi-class settings. The authors conclude that the approach demonstrates practical utility for clinical deployment and underscores the need for domain robustness.

Significance. If the empirical results hold under proper statistical validation, the work supplies a reproducible benchmark for inductive few-shot classification of rare skin conditions using lightweight CNN extractors. The systematic comparison across shot settings and the cross-dataset analysis usefully illustrate domain-shift challenges in medical FSL. Credit is due for the consistent experimental protocol, reliance on public datasets, and the non-parametric classifier that avoids additional trainable parameters.

major comments (2)
  1. [Conclusion] Conclusion: The assertion that the pipeline demonstrates 'practical utility ... for reliable real-world clinical deployment' rests on an untested assumption. The cross-dataset results on MSLD v1.0, MSID, and MSLD v2.0 are presented without any analysis or evidence that these datasets capture clinical variability in skin pigmentation, lighting conditions, camera types, or patient demographics; the observed multi-class degradation could therefore reflect dataset-specific artifacts rather than intrinsic limitations of the method.
  2. [Results] Results section: The central claim that 'MobileNetV2_100 consistently achieves the highest accuracy' is not accompanied by the actual accuracy figures, standard deviations across runs, or any statistical significance tests. Without these quantities it is impossible to verify either the magnitude of the reported superiority or its consistency across the 2/4/6-way and 1/5/10-shot regimes.
minor comments (1)
  1. [Abstract] Abstract: The six CNN backbones are described only as 'widely used' without an explicit list; readers must reach the experimental section to learn the exact architectures and variants (e.g., MobileNetV2_100).

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback, which helps clarify the scope of our claims and the presentation of results. We address each major comment below and will incorporate revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Conclusion] Conclusion: The assertion that the pipeline demonstrates 'practical utility ... for reliable real-world clinical deployment' rests on an untested assumption. The cross-dataset results on MSLD v1.0, MSID, and MSLD v2.0 are presented without any analysis or evidence that these datasets capture clinical variability in skin pigmentation, lighting conditions, camera types, or patient demographics; the observed multi-class degradation could therefore reflect dataset-specific artifacts rather than intrinsic limitations of the method.

    Authors: We agree that the original conclusion overstates the immediate clinical readiness. The three public datasets represent the best available resources for this emerging task, but they are limited in diversity. Our cross-dataset transfer results already quantify substantial degradation in multi-class settings, which we interpret as evidence of domain shift. In revision, we will rephrase the conclusion to state that the pipeline shows practical utility for few-shot classification on available public data while explicitly calling for future work on domain robustness, diverse demographics, and real-world validation before clinical deployment. revision: yes

  2. Referee: [Results] Results section: The central claim that 'MobileNetV2_100 consistently achieves the highest accuracy' is not accompanied by the actual accuracy figures, standard deviations across runs, or any statistical significance tests. Without these quantities it is impossible to verify either the magnitude of the reported superiority or its consistency across the 2/4/6-way and 1/5/10-shot regimes.

    Authors: Tables 2–4 in the manuscript report the per-backbone accuracies for every 2/4/6-way and 1/5/10-shot configuration. However, we acknowledge that these tables currently lack standard deviations (computed over multiple random seeds) and formal statistical tests. We will revise the Results section to include mean accuracies ± standard deviation over five independent runs and add paired t-test p-values comparing MobileNetV2_100 against the other backbones, thereby allowing readers to assess both magnitude and consistency of the observed superiority. revision: yes

Circularity Check

0 steps flagged

No circularity: purely empirical benchmarking with no derivation chain

full rationale

The paper describes an experimental pipeline that extracts features from frozen ImageNet-pretrained CNN backbones and applies the non-parametric SimpleShot classifier on three public datasets across various shot/way settings. All reported results are direct accuracy measurements from these benchmarks and cross-dataset transfers; no equations, fitted parameters, predictions, or uniqueness theorems are invoked. The central claims rest on observed performance numbers rather than any self-referential reduction or imported ansatz, rendering the work self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Central claim rests on transferability of frozen CNN features to this medical domain and on the representativeness of the chosen public datasets; no new entities postulated.

axioms (1)
  • domain assumption Pretrained CNN backbones produce useful embeddings for skin lesion classification without any fine-tuning or adaptation
    Pipeline explicitly freezes the backbone and relies on its off-the-shelf features for all experiments.

pith-pipeline@v0.9.0 · 5582 in / 1212 out tokens · 82474 ms · 2026-05-08T17:41:42.443332+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

56 extracted references · 56 canonical work pages

  1. [1]

    In: 2024 27th International Conference on Computer and Information Technology (ICCIT)

    Abrar, A., Tabassum, F., Ahmed, S.: Performance evaluation of large language models in bangla consumer health query summarization. In: 2024 27th International Conference on Computer and Information Technology (ICCIT). pp. 2748–2753 (2024) 2

    work page 2024

  2. [2]

    In: 2024 6th International Conference on Sustainable Technologies for Industry 5.0 (STI)

    Ahamed, M., Kabir, R.B., Dipto, T.T., Al Mushabbir, M., Ahmed, S., Kabir, M.H.: Perfor- mance analysis of few-shot learning approaches for bangla handwritten character and digit recognition. In: 2024 6th International Conference on Sustainable Technologies for Industry 5.0 (STI). pp. 1–6 (2024) 2

    work page 2024

  3. [3]

    Ahmed, S., Hasan, M.B., Ahmed, T., Kabir, M.H.: DExNet: Combining observations of do- main adapted critics for leaf disease classification with limited data (2025) 2

    work page 2025

  4. [4]

    IEEE Access10, 68868–68884 (2022) 2

    Ahmed, S., Hasan, M.B., Ahmed, T., Sony, M.R.K., Kabir, M.H.: Less is more: Lighter and faster deep neural architecture for tomato leaf disease classification. IEEE Access10, 68868–68884 (2022) 2

    work page 2022

  5. [5]

    In: 2024 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)

    Ahmed, T., Hasan, M.B., Ahmed, S., Kabir, M.H.: ExE-Net: Explainable ensemble network for potato leaf disease classification. In: 2024 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE). pp. 335–339 (2024) 2

    work page 2024

  6. [6]

    Natural Language Processing Journal7, 100079 (2024) 2

    Ahmed, T., Ivan, S., Munir, A., Ahmed, S.: Decoding depression: Analyzing social network insights for depression severity assessment with transformers and explainable ai. Natural Language Processing Journal7, 100079 (2024) 2

    work page 2024

  7. [7]

    In: Proceedings of the 2023 6th International Conference on Machine Vision and Applications

    Ahmed, T., Munir, A., Ahmed, S., Hasan, M.B., Reza, M.T., Kabir, M.H.: Structure- enhanced translation from pet to ct modality with paired gans. In: Proceedings of the 2023 6th International Conference on Machine Vision and Applications. p. 142–146. ICMV A ’23, Association for Computing Machinery, New York, NY , USA (2023) 2

    work page 2023

  8. [8]

    Clinical Immunology (Orlando, Fla.)243, 109108–109108 (2022) 2

    Al-Musa, A., Chou, J., LaBere, B.: The resurgence of a neglected orthopoxvirus: Immuno- logic and clinical aspects of monkeypox virus infections over the past six decades. Clinical Immunology (Orlando, Fla.)243, 109108–109108 (2022) 2

    work page 2022

  9. [9]

    Ali, S.N., Ahmed, M.T., Jahan, T., Paul, J., Sani, S.M.S., Noor, N., Asma, A.N., Hasan, T.: A web-based mpox skin lesion detection system using state-of-the-art deep learning models considering racial diversity (2023) 2, 3, 4, 5, 6

    work page 2023

  10. [10]

    Ali, S.N., Ahmed, M.T., Paul, J., Jahan, T., Sani, S.M.S., Noor, N., Hasan, T.: Monkeypox skin lesion detection using deep learning models: A feasibility study (2022) 2, 3, 4, 5

    work page 2022

  11. [11]

    Arif, N.H., Rabby, S., Papon, M.H.H., Ahmed, S.: Preemptive hallucination reduction: An input-level approach for multimodal language model (2025) 2

    work page 2025

  12. [12]

    IEEE Access10, 92597–92632 (2022) 2

    Ashikur Rahman, A.B.M., Hasan, M.B., Ahmed, S., Ahmed, T., Ashmafee, M.H., Kabir, M.R., Kabir, M.H.: Two decades of bengali handwritten digit recognition: A survey. IEEE Access10, 92597–92632 (2022) 2

    work page 2022

  13. [13]

    In: 2023 26th International Con- ference on Computer and Information Technology (ICCIT)

    Aziz, S., Arif, N.H., Ahbab, S., Ahmed, S., Ahmed, T., Kabir, M.H.: Improved speech emo- tion recognition in bengali language using deep learning. In: 2023 26th International Con- ference on Computer and Information Technology (ICCIT). pp. 1–6 (2023) 2

    work page 2023

  14. [14]

    Neural Networks161(2023) 2, 3, 4, 5, 6

    Bala, D., Hossain, M.S., Hossain, M.A., Abdullah, M.I., Rahman, M.M., Manavalan, B., Gu, N., Islam, M.S., Huang, Z.: Monkeynet: A robust deep convolutional neural network for monkeypox disease detection and classification. Neural Networks161(2023) 2, 3, 4, 5, 6

    work page 2023

  15. [15]

    In: Farkaš, I., Masulli, P., Wermter, S

    Cai, A., Hu, W., Zheng, J.: Few-shot learning for medical image classification. In: Farkaš, I., Masulli, P., Wermter, S. (eds.) Artificial Neural Networks and Machine Learning – ICANN

    work page

  16. [16]

    pp. 441–452. Springer International Publishing, Cham (2020) 2

    work page 2020

  17. [17]

    Medical image analysis 75, 102305 (2021) 2 13

    Cassidy, B., Kendrick, C., Brodzicki, A., Jaworek-Korjakowska, J., Yap, M.H.: Analysis of the isic image datasets: Usage, benchmarks and recommendations. Medical image analysis 75, 102305 (2021) 2 13

    work page 2021

  18. [18]

    Journal of biomedical informatics p

    Chen, B., Han, Y ., Yan, L.: A few-shot learning approach for monkeypox recognition from a cross-domain perspective. Journal of biomedical informatics p. 104449 (2023) 3

    work page 2023

  19. [19]

    Pattern Recognition (2023) 2

    Dai, Z., Yi, J., Yan, L., Xu, Q., Hu, L., Zhang, Q., Li, J., Wang, G.: Pfemed: Few-shot medical image classification using prior guided feature enhancement. Pattern Recognition (2023) 2

    work page 2023

  20. [20]

    Frontiers in Medicine11(2024) 2

    Deng, J., Liu, J., Kong, C., Zang, B., Hu, Y ., Zou, M.: Using novel deep learning models for rapid and efficient assistance in monkeypox screening from skin images. Frontiers in Medicine11(2024) 2

    work page 2024

  21. [21]

    Fuad, T.R., Ahmed, S., Ivan, S.: Aqua20: A benchmark dataset for underwater species clas- sification under challenging conditions (2025) 2

    work page 2025

  22. [22]

    ArXivabs/2211.15459(2022) 2

    Haque, M.E., Ahmed, M.R., Nila, R., Islam, S.: Classification of human monkeypox disease using deep learning models and attention mechanisms. ArXivabs/2211.15459(2022) 2

    work page arXiv 2022

  23. [23]

    Journal of King Saud University - Com- puter and Information Sciences35(7), 101641 (2023) 2

    Hasan, M.B., Ahmed, T., Ahmed, S., Kabir, M.H.: Gaitgcn++: Improving gcn-based gait recognition with part-wise attention and dropgraph. Journal of King Saud University - Com- puter and Information Sciences35(7), 101641 (2023) 2

    work page 2023

  24. [24]

    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition (2015) 3

    work page 2015

  25. [25]

    Huang, G., Liu, Z., van der Maaten, L., Weinberger, K.Q.: Densely connected convolutional networks (2018) 3

    work page 2018

  26. [26]

    Journal of Artificial Intelligence and Metaheuristics (2023) 2

    Islam, A., G., M., El, E.S.M., Moustafa, H.E.D.: Medical image classification for monkey- pox case using deep learning algorithms: A survey. Journal of Artificial Intelligence and Metaheuristics (2023) 2

    work page 2023

  27. [27]

    In: 2024 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE)

    Ivan, S., Ahmed, T., Ahmed, S., Kabir, M.H.: A vision-language multimodal framework for detecting hate speech in memes. In: 2024 IEEE Canadian Conference on Electrical and Computer Engineering (CCECE). pp. 464–468 (2024) 2

    work page 2024

  28. [28]

    IEEE Journal of Biomedical and Health Informat- ics27, 17–28 (2022) 2

    Jiang, H., Gao, M., Li, H., Jin, R., Miao, H., Liu, J.: Multi-learner based deep meta-learning for few-shot medical image classification. IEEE Journal of Biomedical and Health Informat- ics27, 17–28 (2022) 2

    work page 2022

  29. [29]

    In: 2023 International Joint Conference on Neural Networks (IJCNN)

    Khan, A.M., Ashrafee, A., Khan, F.S., Hasan, M.B., Kabir, M.H.: Attresdu-net: Medical image segmentation using attention-based residual double u-net. In: 2023 International Joint Conference on Neural Networks (IJCNN). pp. 1–8 (2023) 2

    work page 2023

  30. [30]

    In: 25th International Conference on Computer and Infor- mation Technology (ICCIT)

    Khan, A.M., Ashrafee, A., Sayera, R., Ivan, S., Ahmed, S.: Rethinking cooking state recog- nition with vision transformers. In: 25th International Conference on Computer and Infor- mation Technology (ICCIT). pp. 170–175 (2022) 2

    work page 2022

  31. [31]

    In: Proceedings of the First Workshop on Bangla Language Processing (BLP-2023)

    Khan, A., Kamal, F., Chowdhury, M.A., Ahmed, T., Laskar, M.T.R., Ahmed, S.: BanglaCHQ-summ: An abstractive summarization dataset for medical queries in Bangla con- versational speech. In: Proceedings of the First Workshop on Bangla Language Processing (BLP-2023). pp. 85–93. Association for Computational Linguistics, Singapore (Dec 2023) 2

    work page 2023

  32. [32]

    In: Findings of the Association for Computational Linguistics: EMNLP 2023

    Khan, A., Kamal, F., Nower, N., Ahmed, T., Ahmed, S., Chowdhury, T.: NERvous about my health: Constructing a Bengali medical named entity recognition dataset. In: Findings of the Association for Computational Linguistics: EMNLP 2023. pp. 5768–5774. Association for Computational Linguistics, Singapore (Dec 2023) 2

    work page 2023

  33. [33]

    Journal of Innovative Image Processing (2025) 2

    Kottath, A.V ., P, R.: Poxtlnet50: Deep learning-based approach for accurate image detection of monkeypox disease. Journal of Innovative Image Processing (2025) 2

    work page 2025

  34. [34]

    IEEE Access12, 32819–32829 (2024) 2

    Kundu, D., Rahman, M.M., Rahman, A., Das, D., Siddiqi, U.R., Alam, M.G.R., Dey, S.K., Muhammad, G., Ali, Z.: Federated deep learning for monkeypox disease detection on gan- augmented dataset. IEEE Access12, 32819–32829 (2024) 2

    work page 2024

  35. [35]

    IEEE Transactions on Geoscience and Remote Sensing57, 2290–2304 (2019) 2

    Liu, B., Yu, X., Yu, A., Zhang, P., Wan, G., Wang, R.: Deep few-shot learning for hyper- spectral image classification. IEEE Transactions on Geoscience and Remote Sensing57, 2290–2304 (2019) 2

    work page 2019

  36. [36]

    In: Proceedings of the Conference on Empirical Methods in Natural Language Processing

    Mahbub, R., Khan, I., Anuva, S., Shahriar, M.S., Laskar, M.T.R., Ahmed, S.: Unveiling the essence of poetry: Introducing a comprehensive dataset and benchmark for poem sum- 14 marization. In: Proceedings of the Conference on Empirical Methods in Natural Language Processing. pp. 14878–14886. Association for Computational Linguistics (2023) 2

    work page 2023

  37. [37]

    Morbidity and Mortality Weekly Report72, 68–72 (2023) 2

    McCollum, A., Shelus, V .S., Hill, A., Traore, T., Onoja, B., Nakazawa, Y ., Doty, J.B., Yinka- Ogunleye, A., Petersen, B., Hutson, C., Lewis, R.F.: Epidemiology of human mpox — world- wide, 2018–2021. Morbidity and Mortality Weekly Report72, 68–72 (2023) 2

    work page 2018

  38. [38]

    Lancet (London, England)401, 60–74 (2022) 1

    Mitjà, O., Ogoina, D., Titanji, B.K., Galván, C., Muyembe, J.J., Marks, M., Orkin, C.: Mon- keypox. Lancet (London, England)401, 60–74 (2022) 1

    work page 2022

  39. [39]

    Cureus15(2023) 2

    Nisar, H., Saleem, O., Sapna, F., Sham, S., Perkash, R.S., Kiran, N., Anjali, F., Mehreen, A., Ram, B.: A narrative review on the monkeypox virus: An ongoing global outbreak hitting the non-endemic countries. Cureus15(2023) 2

    work page 2023

  40. [40]

    IEEE Transactions on Medical Imaging41(2022) 2

    Ouyang, C., Biffi, C., Chen, C., Kart, T., Qiu, H., Rueckert, D.: Self-supervised learning for few-shot medical image segmentation. IEEE Transactions on Medical Imaging41(2022) 2

    work page 2022

  41. [41]

    Medical Science Monitor : International Medical Journal of Experimental and Clinical Research28, e938203–1 – e938203–3 (2022) 2

    Parums, D.: Editorial: Current status of non-endemic global infections with the monkeypox virus. Medical Science Monitor : International Medical Journal of Experimental and Clinical Research28, e938203–1 – e938203–3 (2022) 2

    work page 2022

  42. [42]

    GeroScience44(2022) 2

    Quarleri, J., Delpino, M., Galvan, V .: Monkeypox: considerations for the understanding and containment of the current outbreak in non-endemic countries. GeroScience44(2022) 2

    work page 2022

  43. [43]

    In: ICCV 2025 Workshop on Cultural Continuity of Artists (2025) 2

    Raiyan, S.R., Amio, Z.Z., Ahmed, S.: HaSPeR: An image repository for hand shadow puppet recognition. In: ICCV 2025 Workshop on Cultural Continuity of Artists (2025) 2

    work page 2025

  44. [44]

    Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., Chen, L.C.: Mobilenetv2: Inverted resid- uals and linear bottlenecks (2019) 3

    work page 2019

  45. [45]

    histopatho- logical images (2024) 8

    Sekhar, A., Bhattacharya, A., Goyal, V ., Goel, V ., Bhangale, A., Gupta, R.K., Sethi, A.: Cross-domain evaluation of few-shot classification models: Natural images vs. histopatho- logical images (2024) 8

    work page 2024

  46. [46]

    Infection50, 1425–1430 (2022) 2

    Shafaati, M., Zandi, M.: State-of-the-art on monkeypox virus: an emerging zoonotic disease. Infection50, 1425–1430 (2022) 2

    work page 2022

  47. [47]

    Simonyan, K., Zisserman, A.: Very deep convolutional networks for large-scale image recog- nition (2015) 3

    work page 2015

  48. [48]

    Pattern Recognition120, 108111 (2021) 2

    Singh, R., Bharti, V ., Purohit, V ., Kumar, A., Singh, A.K., Singh, S.K.: Metamed: Few-shot medical image classification using gradient-based meta-learning. Pattern Recognition120, 108111 (2021) 2

    work page 2021

  49. [49]

    Szegedy, C., Vanhoucke, V ., Ioffe, S., Shlens, J., Wojna, Z.: Rethinking the inception archi- tecture for computer vision (2015) 3

    work page 2015

  50. [50]

    Tabassum, F., Islam, S., Rizwan, S., Sobhan, M., Ahmed, T., Ahmed, S., Chowdhury, T.M.: Precision cancer classification and biomarker identification from mrna gene expression via dimensionality reduction and explainable ai (2024) 2

    work page 2024

  51. [51]

    Tan, M., Le, Q.V .: Efficientnet: Rethinking model scaling for convolutional neural networks (2020) 3

    work page 2020

  52. [52]

    ACM Computing Surveys 55, 1–40 (2022) 2

    Wang, T.Y ., Song, Y ., Cai, P., Sahoo, J.P., Mondal, S.: A comprehensive survey of few-shot learning: Evolution, applications, challenges, and opportunities. ACM Computing Surveys 55, 1–40 (2022) 2

    work page 2022

  53. [53]

    Learning robust global representations by penalizing local predictive power

    Wang, Y ., Chao, W.L., Weinberger, K.Q., van der Maaten, L.: Simpleshot: Revisiting nearest- neighbor classification for few-shot learning. arXiv:1911.04623 (2019) 3, 4, 8

    work page arXiv 1911

  54. [54]

    ArXivabs/1904.05046(2019) 2

    Wang, Y ., Yao, Q.: Few-shot learning: A survey. ArXivabs/1904.05046(2019) 2

    work page arXiv 1904

  55. [55]

    In: 2021 Joint 10th International Confer- ence on Informatics, Electronics & Vision (ICIEV) and 2021 5th International Conference on Imaging, Vision & Pattern Recognition (icIVPR)

    Yasmeen, A., Rahman, F.I., Ahmed, S., Kabir, M.H.: Csvc-net: Code-switched voice com- mand classification using deep cnn-lstm network. In: 2021 Joint 10th International Confer- ence on Informatics, Electronics & Vision (ICIEV) and 2021 5th International Conference on Imaging, Vision & Pattern Recognition (icIVPR). pp. 1–8 (2021) 2

    work page 2021

  56. [56]

    IET Image Process.17, 3589–3598 (2023) 2 15

    Zhang, B., Gao, B., Liang, S., Li, X., Wang, H.: A classification algorithm based on improved meta learning and transfer learning for few-shot medical images. IET Image Process.17, 3589–3598 (2023) 2 15

    work page 2023