Pith. sign in

REVIEW 5 major objections 6 minor 48 references

An empirical study for the early detection of Mpox from skin lesion images using pretrained CNN models leveraging XAI technique

T0 review · 5 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read InceptionV3 hits 95% on Mpox images; MobileNetV2 93% on six classes

desk verdict Routine transfer-learning benchmark whose headline accuracies are unreliable due to ambiguous augmentation/split order and internally inconsistent AUC metrics. read the letter →

arxiv 2507.15915 v1 pith:6YLW2QEZ submitted 2025-07-21 cs.CV

Mohammad Asifur Rahim , Muhammad Nazmul Arefin , Md. Mizanur Rahman , Md Ali Hossain , Ahmed Moustafa This is my paper · ORCID
classification cs.CV
keywords MpoxdetectiontransferlearningpretrainedCNNInceptionV3MobileNetV2Grad-CAMexplainableAIskinlesionclassification
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper claims that fine-tuning four ImageNet-pretrained convolutional networks—VGG16, VGG19, InceptionV3, and MobileNetV2—produces accurate classifiers for Mpox from skin-lesion photos, with InceptionV3 reaching 95% accuracy on a two-class Kaggle dataset and MobileNetV2 reaching 93% on a six-class dataset. The authors argue this is evidence that lightweight transfer learning can support early detection of a disease that is easy to confuse with chickenpox, measles, and other rashes. They also claim Grad-CAM heatmaps highlight the lesion regions behind each prediction, giving clinicians a visual reason to trust or question the output. Because the study validates on both binary and multi-class data, the authors present the results as a step toward deployable, interpretable screening tools.

What carries the argument

The mechanism is transfer learning from ImageNet: every pretrained convolutional base is frozen, and only a custom classification head is trained. The head consists of a flatten layer, a dense layer with 256 ReLU units, a dropout layer at rate 0.5, and a softmax output; early stopping and model checkpointing cap training at 30 epochs. Grad-CAM then computes gradient-weighted activations from the final convolutional layer to produce a heatmap of the image regions that most influenced the prediction. This combination is what the paper claims makes the models lightweight, fast, and interpretable.

What would settle it

Rerun the same four-model pipeline on the two Kaggle datasets with augmentation applied strictly after the train/validation/test split, and report accuracy on the untouched original test images; if the figures fall materially below 95% on binary and 93% on six-class data, the reported numbers were inflated by leakage of augmented copies into training.

Watch

Extended reading notes

Core claim

On the paper's own terms, the core discovery is that a frozen pretrained feature extractor plus a small custom head—a flatten layer, a dense layer of 256 ReLU units, a dropout layer at 0.5, and a softmax output—is enough to reach near-clinical accuracy on both datasets. InceptionV3 achieves 0.95 accuracy, precision, recall, and F1 on the binary Monkeypox Skin Lesion Dataset, and MobileNetV2 achieves 0.93 accuracy and precision with 0.90 recall on the six-class MSLD v2.0. The paper further reports that freezing the ImageNet base shrinks trainable parameters substantially, with VGG16 dropping from about 21 million total parameters to about 6.4 million trainable, and that Grad-CAM highlights regions corresponding to lesion areas. The stated conclusion is that these lightweight, interpretable models generalize across two independent datasets and are competitive with prior state-of-the-art results.

Load-bearing premise

The reported accuracies are only meaningful as generalization estimates if augmented copies of the same original photograph never appear in both the training and test sets, and the paper does not show that the data split was performed before augmentation.

Editorial extensions

If this is right

  • If correct, InceptionV3 provides a binary Mpox-versus-others screen at 95% accuracy and MobileNetV2 a six-class differential at 93% accuracy on these Kaggle datasets.
  • Per-class AUC values up to 0.99 on binary and 1.00 on some multi-class categories indicate these models separate Mpox from common look-alikes on the images tested.
  • Grad-CAM heatmaps give a clinician a visual region to check, which is the core of the paper's interpretability claim.
  • Wall times of about 16 minutes for InceptionV3 on binary and 57 minutes for MobileNetV2 on multi-class suggest the fine-tuned heads can be retrained on moderate hardware.
  • The reduced trainable parameter counts, such as VGG16 shrinking from about 21M to 6.4M parameters, support deployment on resource-limited devices.

Reading between the lines

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

  • The paper's own future-work statement admits the link between Grad-CAM heatmaps and model performance was not explicitly established, so the heatmaps should be treated as illustrative rather than validated clinical evidence.
  • If the augmentation-before-split pipeline inflated the reported accuracies, the true ranking of the four models on unseen patient images is unknown and could differ from InceptionV3/MobileNetV2 leading.
  • A direct deployment test would classify only the original, non-augmented images, one per patient, to measure the accuracy a clinician would actually experience.
  • The method's applicability to other rash-like diseases could be tested by fine-tuning the same frozen bases on an independent dermatology dataset with a held-out patient split.
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 presents an empirical benchmark of four pretrained CNN architectures (VGG16, VGG19, InceptionV3, MobileNetV2) fine-tuned with a small custom head for early detection of mpox from skin lesion images. Two public Kaggle datasets are used: a two-class MSLD dataset and a six-class MSLD v2.0 dataset. The authors report that InceptionV3 achieves 95% accuracy on the binary dataset and MobileNetV2 achieves 93% accuracy on the multi-class dataset, and they use Grad-CAM to visualize which image regions drive predictions. The paper also reports precision, recall, F1-score, loss, wall time, and CPU time, and compares the results with selected prior work.

Significance. If the reported numbers were reliable, this would be a useful but incremental empirical confirmation that standard transfer learning with frozen pretrained backbones can classify skin-lesion images on these specific public datasets. The comparison across four architectures, the reporting of wall/CPU time, and the use of two datasets are positive features, as is the candid discussion of some limitations in Sections 6 and 7. However, the contribution is primarily an evaluation rather than a methodological advance, and the paper provides no code or evaluation scripts to support reproducibility. The significance is therefore conditional on the trustworthiness of the reported metrics, and the concerns below put that trustworthiness in question.

major comments (5)
  1. [§3.1 (Figure 3), §3.2, §3.3.2] The reported test accuracies are intended to measure generalization to unseen images, but the pipeline described in Figure 3 and Section 3.3.2 applies data augmentation before the train/validation/test split, and Section 3.2 states that only the pre-augmented Kaggle versions are used. Consequently, augmented copies of the same original lesion image can appear in both the training and test partitions, and the manuscript does not state that the split was performed on original-image identity (nor is code provided to rule this out). This directly threatens the central claims of 0.95 accuracy for InceptionV3 in Table 8 and 0.93 accuracy for MobileNetV2 in Table 9. The authors must re-run the experiments with augmentation contained within the training split, or provide evidence that no source-image overlap exists across splits.
  2. [§4.1, §4.2 (ROC analysis)] The micro-average AUC values are mathematically inconsistent with the other reported metrics. In a binary problem the micro-average ROC is identical to the overall ROC, so values of 0.33–0.39 for models with per-class AUCs of 0.95–0.99 cannot arise from a correct ROC computation. In the multi-class setting, micro-average AUCs of 0.37–0.61 alongside accuracies of 0.85–0.93 and per-class AUCs of 0.93–1.00 are likewise implausible. This indicates a systematic error in the ROC/AUC computation and raises doubt about all metrics produced by the same evaluation pipeline.
  3. [§3.3.7 (Eq. 1)] The paper's metric definitions are incorrect: Eq. (1) defines accuracy as (TP+FP)/Total instead of (TP+TN)/Total, and the surrounding text defines FP as "Actual Positive but predicted negative" and FN as "Actual negative but predicted positive", interchanging the two. Because accuracy, precision, recall, and F1-score are the load-bearing outcomes of the study, the authors need to correct these definitions and confirm that the reported numbers in Tables 8 and 9 were computed with a standard implementation.
  4. [§7 (Future work) and §5 (Discussion)] The authors themselves state in Section 7 that "the relationship between Grad-CAM outputs and the CNN model's performance was not explicitly established, leaving the relevance of CNN predictions and Grad-CAM visualizations unclear." This admission undercuts the interpretability component of the contribution: Section 5 describes the Grad-CAM heatmaps as validating "the clinical relevance of predictions," which is stronger than the evidence supports. The paper should either provide a quantitative or structured evaluation linking Grad-CAM localizations to model decisions, or substantially weaken the claim. If this is not done, the XAI claim should be treated as illustrative rather than validated.
  5. [§3.3.2, Tables 4–5] All results are reported from a single 75/15/10 split, with no repeated runs, cross-validation, confidence intervals, or significance tests. Given the small number of original images (roughly 400 and 1000 before augmentation, as noted in Section 3.2 and Table 3), the 0.93–0.95 accuracies could vary substantially across splits. The authors should report variability, for example through repeated stratified splits or k-fold cross-validation, before claiming that one architecture demonstrated the best performance.
minor comments (6)
  1. [§3.3.6, Table 6, §8] The learning rate is stated as 0.001 in the text of Section 3.3.6 but as 0.0001 in Table 6, and the Conclusion repeats 0.0001; these values should be reconciled.
  2. [Table 3 vs Tables 1–2] Table 3 reports roughly 1,200 augmented images for MSLD and more than 3,000 for MSLD v2.0, but Tables 1 and 2 and the text in Section 3.2 use 3,192 and 7,532 images for the augmented versions; the counts should be made consistent.
  3. [Tables 8 and 11] In Section 4.1, the CPU time for InceptionV3 is reported as "3m 60s" in Table 8, while Table 11 lists 3m 16s for the same model; the entry should be corrected, and "3m 60s" should be written as 4m.
  4. [§3.1.2, §5] Several cross-references are inaccurate: Section 5 refers to confusion matrices in "figure 7 and figure 10," and the research-question roadmap in Section 3.1.2 points to sections that do not match the actual numbering; these need to be updated.
  5. [Throughout] The manuscript contains many typographical and grammatical errors (for example, "this models were basically trained," "videnced," "Orthodox virus genus" for Orthopoxvirus, and "Figure 5 and 6 displays") that should be corrected by a thorough language edit.
  6. [§9] No code or training/evaluation script is provided, and the only availability statement is a Kaggle link; providing the split and evaluation code would materially improve reproducibility and would help resolve the leakage concern raised in the first major comment.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the empirical benchmark is measured on external public datasets, with no fitted parameter, derivation chain, or load-bearing self-citation reused as evidence.

full rationale

This paper is an empirical benchmark, not a derivation. It fine-tunes four pretrained CNN models (VGG16, VGG19, InceptionV3, MobileNetV2) on two public Kaggle datasets and reports measured accuracy, precision, recall, F1, loss, and Grad-CAM visualizations. No equation or construction step equates an output to an input: the reported 95% and 93% accuracies are empirical measurements made on held-out portions of those external datasets, and no parameter is fitted to test outcomes and then renamed as a prediction. The dataset sources cited as [49] and [50] are works by other research groups, not the present authors, so the paper does not lean on a load-bearing self-citation chain. Likewise, no ‘uniqueness theorem’ from the authors' prior work is imported, and the transfer-learning procedure and Grad-CAM technique are standard external methods. The paper does contain serious validity threats: Figure 3 and Section 3.3.2 describe augmentation being applied before the train/validation/test split, which can allow augmented copies of the same source image to appear in both training and test partitions, plausibly inflating the reported accuracies; the micro-average AUC values of 0.33–0.39 for binary models with per-class AUCs of 0.95–0.99 are also internally inconsistent. However, data leakage and metric inconsistencies are not circularity: they concern whether the test set measures generalization, not whether the paper's conclusions reduce by construction to its own inputs. The paper itself admits in Section 7 that the relationship between Grad-CAM outputs and model performance was not explicitly established, which is a limitation statement but again not a circular step. There is no fitted parameter reused as evidence, no self-citation carrying the central claim, and no definitional equivalence between input and output. Accordingly, the circularity burden is minimal and the honest finding is 'no significant circularity' with score 0.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The paper introduces no new entities and no derivation. Its results rest on standard pretrained architectures, two public datasets, hand-picked hyperparameters, and the unstated assumption that augmentation does not leak across the split.

free parameters (7)
  • Dense layer size (256) = 256
    Chosen by hand as the width of the single hidden layer added on top of the frozen base model; no ablation or search is reported.
  • Dropout rate (0.5) = 0.5
    Chosen to reduce overfitting; no sensitivity analysis is reported.
  • Learning rate = 0.0001 (Table 6) or 0.001 (Section 3.3.6 text)
    The paper states both values in different places, an internal inconsistency; chosen by hand.
  • Batch size (16) = 16
    Chosen for memory and training stability; not tuned.
  • Early stopping patience (5) = 5
    Chosen by hand to balance training time and performance.
  • Train/validation/test split ratio (75/15/10) = 75/15/10
    Chosen by the authors; no cross-validation or repeated splits are used.
  • MSLD v2.0 fold selection = fifth fold of the augmented dataset
    The authors used only the fifth fold of the augmented MSLD v2.0 to save computational overhead, a manual subset selection that can bias results.
assumptions (4)
  • domain assumption ImageNet-pretrained features transfer to skin lesion classification
    The method freezes ImageNet weights and adds a small head; the paper itself notes in Threats to Validity that ImageNet features may not align with medical imaging.
  • domain assumption The two Kaggle datasets are correctly labeled and representative
    The evaluation relies on the labels from Ali et al. for MSLD and MSLD v2.0; no clinician validation of labels is reported.
  • domain assumption Data augmentation preserves labels and does not leak information across the train/test split
    The pipeline augments images and then splits into train/val/test, so augmented copies of the same source image can appear in both training and test sets; the paper never states that augmentation was applied after splitting.
  • domain assumption Grad-CAM heatmaps correspond to clinically meaningful features
    The paper treats Grad-CAM highlights as evidence of trustworthiness, but in Future Work it admits the relationship between Grad-CAM outputs and CNN performance was not explicitly established.

how reviews work

0 comments
Cite this review

Pith. "Pith review of An empirical study for the early detection of Mpox from skin lesion images using pretrained CNN models leveraging XAI technique." pith.science (2026). https://pith.science/paper/6YLW2QEZ

@misc{pith2026250715915,
  author       = {Pith},
  title        = {Pith review of: An empirical study for the early detection of Mpox from skin lesion images using pretrained CNN models leveraging XAI technique},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6YLW2QEZ}},
  note         = {Machine review of arXiv:2507.15915}
}
read the original abstract

Context: Mpox is a zoonotic disease caused by the Mpox virus, which shares similarities with other skin conditions, making accurate early diagnosis challenging. Artificial intelligence (AI), especially Deep Learning (DL), has a strong tool for medical image analysis; however, pre-trained models like CNNs and XAI techniques for mpox detection is underexplored. Objective: This study aims to evaluate the effectiveness of pre-trained CNN models (VGG16, VGG19, InceptionV3, MobileNetV2) for the early detection of monkeypox using binary and multi-class datasets. It also seeks to enhance model interpretability using Grad-CAM an XAI technique. Method: Two datasets, MSLD and MSLD v2.0, were used for training and validation. Transfer learning techniques were applied to fine-tune pre-trained CNN models by freezing initial layers and adding custom layers for adapting the final features for mpox detection task and avoid overfitting. Models performance were evaluated using metrics such as accuracy, precision, recall, F1-score and ROC. Grad-CAM was utilized for visualizing critical features. Results: InceptionV3 demonstrated the best performance on the binary dataset with an accuracy of 95%, while MobileNetV2 outperformed on the multi-class dataset with an accuracy of 93%. Grad-CAM successfully highlighted key image regions. Despite high accuracy, some models showed overfitting tendencies, as videnced by discrepancies between training and validation losses. Conclusion: This study underscores the potential of pre-trained CNN models in monkeypox detection and the value of XAI techniques. Future work should address dataset limitations, incorporate multimodal data, and explore additional interpretability techniques to improve diagnostic reliability and model transparency

Figures

Figures reproduced from arXiv: 2507.15915 by the authors.

Figure 1
Figure 1. Detailed methodology of our work [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Model Architecture arXiv:2507.15915v1 [cs.CV] 21 Jul 2025 [PITH_FULL_IMAGE:figures/full_fig_p001_2.png] view at source ↗
Figure 3
Figure 3. Detailed methodology of our work resizing and normalization. Subsequently, data augmentation techniques were applied to increase the number of images and enhance diversity. The datasets were then split into training, validation, and test sets. Four pre￾trained models—VGG16, VGG19, InceptionV3, and MobileNetV2—were fine￾tuned by integrating our proposed custom layers. The performance of these models was evaluated on … view at source ↗
Figures from the paper (13 more)
Figure 4
Figure 4. Figure 4: Race wise Mpox distribution offers consistent image resolution, detailed annotations, and metadata that facilitate advanced analysis and model training. Its high diversity makes it more suitable for developing models with better generalization capabilities [PITH_FULL_…
Figure 5
Figure 5. Figure 5: Mpox Skin Lesion Dataset V2 (MSLD v2.0) (a) Chickenpox, (b) Cowpox, (c) [PITH_FULL_IMAGE:figures/full_fig_p015_5.png]
Figure 6
Figure 6. Figure 6: Monkeypox Skin Lesion Dataset (MSLD) (a) Monkeypox, (b) Others [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 7
Figure 7. Figure 7: Model architecture VGG) and achieved high accuracy on the ILSVRC dataset. MobileNetV2 MobileNet V2 [38] is a lightweight CNN architecture de￾signed for mobile and embedded vision applications. It is an improved ver￾sion of MobileNet V1 that emphasizes performance for d…
Figure 8
Figure 8. Figure 8: Modified Layer 1. Initially, the base models have been initialized with a pre-trained network that lacks fully connected layers. 2. To preserve the knowledge learned during ImageNet training, all layers of the base model were frozen. This ensured that the feature extra…
Figure 9
Figure 9. Figure 9: Confusion matrices for four different models for binary [PITH_FULL_IMAGE:figures/full_fig_p025_9.png]
Figure 10
Figure 10. Figure 10: Accuracy-loss curve for four different models for binary class [PITH_FULL_IMAGE:figures/full_fig_p026_10.png]
Figure 11
Figure 11. Figure 11: Visual comparison of model performance: ROC curves [PITH_FULL_IMAGE:figures/full_fig_p027_11.png]
Figure 12
Figure 12. Figure 12: Grad-CAM visualizations for the four models for binary class. [PITH_FULL_IMAGE:figures/full_fig_p028_12.png]
Figure 13
Figure 13. Figure 13: Confusion matrices for four different models for multi-class dataset [PITH_FULL_IMAGE:figures/full_fig_p031_13.png]
Figure 14
Figure 14. Figure 14: Accuracy and loss curves for all the four models for multi-class dataset [PITH_FULL_IMAGE:figures/full_fig_p032_14.png]
Figure 15
Figure 15. Figure 15: Visual comparison of model performance: ROC curves [PITH_FULL_IMAGE:figures/full_fig_p033_15.png]
Figure 16
Figure 16. Figure 16: Grad-CAM visualizations for the four models for multi class. [PITH_FULL_IMAGE:figures/full_fig_p034_16.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 35 canonical work pages

  1. [1]

    Epidemiologi- cal situation of mon- keypox transmission by possible sexual contact: a systematic review

    Le ´on-Figueroa, D.A., Barboza, J.J., Garcia-Vasquez, E.A., Bonilla- Aldana, D.K., Diaz-Torres, M., Salda˜na-Cumpa, H.M., Diaz-Murillo, M.T., Cruz, O.C.S., Rodriguez-Morales, A.J., 2022. Epidemiologi- cal situation of mon- keypox transmission by possible sexual contact: a systematic review. Trop- ical Medicine and Infectious Disease 7, 267.https://doi.org...

    work page doi:10.3390/tropicalmed7100267 2022

  2. [2]

    Burden of salmonellosis, campylobacteriosis and liste- riosis: a time series analysis, belgium, 2012 to 2020

    de Noordhout, C.M., Devleesschauwer, B., Haagsma, J.A., Havelaar, A.H., Bertrand, S., Vandenberg, O., Quoilin, S., Brandt, P.T., Spey- broeck, N., 2017. Burden of salmonellosis, campylobacteriosis and liste- riosis: a time series analysis, belgium, 2012 to 2020 . Eurosurveillance 22, 30615. https://doi.org/10.2807/1560-7917

    work page doi:10.2807/1560-7917 2017

  3. [3]

    Genomic variability of monkeypox virus among hu- mans, democratic republic of the congo

    Kugelman, J.R., Johnston, S.C., Mulembakani, P.M., Kisalu, N., Lee, M.S., Koroleva, G., McCarthy, S.E., Gestole, M.C., Wolfe, N.D., Fair, J.N., et al., 2014. Genomic variability of monkeypox virus among hu- mans, democratic republic of the congo. Emerging infectious diseases 20, 232

    work page 2014

  4. [4]

    Mon- keypox: a comprehensive review of transmission, pathogenesis, and man- ifestation

    Kaler, J., Hussain, A., Flores, G., Kheiri, S., Desrosiers, D., 2022. Mon- keypox: a comprehensive review of transmission, pathogenesis, and man- ifestation. Cureus 14

    work page 2022

  5. [5]

    Monkeypox virus emerges from the shadow of its more infamous cousin: fam- ily biology matters

    Xiang, Y., White, A., 2022. Monkeypox virus emerges from the shadow of its more infamous cousin: fam- ily biology matters . Emerging microbes & infections 11, 1768–1777.https://doi.org/10.1080/22221751.2022.2095309 39

    arXiv 2022

  6. [6]

    Re-emerging human monkeypox: a major public-health debacle

    Saxena, S.K., Ansari, S., Maurya, V.K., Kumar, S., Jain, A., Paweska, J.T., Tri- pathi, A.K., Abdel-Moneim, A.S., 2023. Re-emerging human monkeypox: a major public-health debacle. Journal of medical virology 95, e27902. https://doi.org/10.1002/jmv.27902

    work page doi:10.1002/jmv.27902 2023

  7. [7]

    Human monkeypox disease (mpx)

    Farahat, R.A., Sah, R., El-Sakka, A.A., Benmelouka, A.Y., Kundu, M., Labieb, F., Shaheen, R.S., Abdelaal, A., Abdelazeem, B., Bonilla- Aldana, D.K., et al., 2022. Human monkeypox disease (mpx) . Le in- fezioni in Medicina 30, 372.https://doi.org/10.53854/liim-3003-6

    work page doi:10.53854/liim-3003-6 2022

  8. [8]

    A global up- date of mpox (monkeypox) in children

    Sam-Agudu, N.A., Martyn-Dickens, C., Ewa, A.U., 2023. A global up- date of mpox (monkeypox) in children. Current Opinion in Pediatrics 35, 193–200

    work page 2023

Show all 48 references
  1. [9]

    Outbreak of human monkeypox in nigeria in 2017–18: a clinical and epidemiological report

    Yinka-Ogunleye, A., Aruna, O., Dalhat, M., Ogoina, D., McCollum, A., Disu, Y., Mamadu, I., Akinpelu, A., Ahmad, A., Burga, J., et al., 2019. Outbreak of human monkeypox in nigeria in 2017–18: a clinical and epidemiological report. The Lancet Infectious Diseases 19, 872–879.htt...

    2019 doi

  2. [10]

    Recent outbreak of monkeypox: implications for pub- lic health recommendations and crisis management in india

    Kumar, A., Borkar, S.K., Choudhari, S.G., Mendhe, H.G., Bankar, N.J., Mendhe, H., 2023. Recent outbreak of monkeypox: implications for pub- lic health recommendations and crisis management in india. Cureus 15

    2023

  3. [11]

    Monkey- pox virus infection in humans across 16 coun- tries—april–june 2022

    Thornhill, J.P., Barkati, S., Walmsley, S., Rockstroh, J., Antinori, A., Harrison, L.B., Palich, R., Nori, A., Reeves, I., Habibi, M.S., et al., 2022. Monkey- pox virus infection in humans across 16 coun- tries—april–june 2022. New England Journal of Medicine 387, 679–691. htt...

    2022 doi

  4. [12]

    Monkeypox virus: a re- emergent threat to humans

    Gong, Q., Wang, C., Chuai, X., Chiu, S., 2022. Monkeypox virus: a re- emergent threat to humans . Virologica Sinica 37, 477–482.https://doi.org/10.1016/j.virs.2022.07.006

    2022 doi

  5. [13]

    Use of artificial intelligence in infectious diseases, in: Artificial intelligence in precision health

    Agrebi, S., Larbi, A., 2020. Use of artificial intelligence in infectious diseases, in: Artificial intelligence in precision health. Elsevier, pp. 415–438.https://doi.org/10.1016/B978-0-12-817133-2.00018-5

    2020 doi

  6. [14]

    Machine-learning-based disease diagnosis: A comprehensive review , in: Healthcare, MDPI

    Ahsan, M.M., Luna, S.A., Siddique, Z., 2022. Machine-learning-based disease diagnosis: A comprehensive review , in: Healthcare, MDPI. p. 541.https://doi.org/10.3390/healthcare10030541 40

    2022 doi

  7. [15]

    Ex- plainable artificial intelligence (xai) in deep learning-based medical image analysis

    Van der Velden, B.H., Kuijf, H.J., Gilhuijs, K.G., Viergever, M.A., 2022. Ex- plainable artificial intelligence (xai) in deep learning-based medical image analysis . Medical Image Analysis 79, 102470.https://doi.org/10.1016/j.media.2022.102470

    2022

  8. [16]

    A cnn-based methodology for breast can- cer diagnosis using thermal images

    Zuluaga-Gomez, J., Al Masry, Z., Benaggoune, K., Meraghni, S., Zerhouni, N., 2021. A cnn-based methodology for breast can- cer diagnosis using thermal images. Computer Methods in Biome- chanics and Biomedical Engineering: Imaging & Visualization 9, 131–145.https://doi.org/10.1...

    2021

  9. [17]

    Medical image analysis using deep learning algorithms

    Li, M., Jiang, Y., Zhang, Y., Zhu, H., 2023. Medical image analysis using deep learning algorithms. Frontiers in public health 11, 1273253. https://doi.org/10.3389/fpubh.2023.1273253

    2023

  10. [18]

    Transfer learning for medical im- age classification: a litera- ture review

    Kim, H.E., Cosa-Linan, A., Santhanam, N., Jannesari, M., Maros, M.E., Gans- landt, T., 2022. Transfer learning for medical im- age classification: a litera- ture review. BMC medical imaging 22, 69.https://doi.org/10.1186/s12880-022-00793-7

    2022 doi

  11. [19]

    Transformers in medical image analysis

    He, K., Gan, C., Li, Z., Rekik, I., Yin, Z., Ji, W., Gao, Y., Wang, Q., Zhang, J., Shen, D., 2023. Transformers in medical image analysis. In- telligent Medicine 3, 59–78.https://doi.org/10.1016/j.imed.2022.07.002

    2023 doi

  12. [21]

    Poxnet22: A fine-tuned model for the classification of monkeypox disease using transfer learning

    Yasmin, F., Hassan, M.M., Hasan, M., Zaman, S., Kaushal, C., El- Shafai, W., Soliman, N.F., 2023. Poxnet22: A fine-tuned model for the classification of monkeypox disease using transfer learning. Ieee Access 11, 24053–24076.https://doi.org/10.1109/ACCESS.2023.3253868

    2023

  13. [22]

    Monkey- pox diagnosis with inter- pretable deep learning

    Ahsan, M.M., Ali, M.S., Hassan, M.M., Abdullah, T.A., Gupta, K.D., Bagci, U., Kaushal, C., Soliman, N.F., 2023. Monkey- pox diagnosis with inter- pretable deep learning. IEEE Access 11, 81965–81980.https://doi.org/10.1016/j.eswa.2022.119483

    2023

  14. [23]

    Attention to monkeypox: An 41 inter- pretable monkeypox detection technique using attention mecha- nism

    Raha, A.D., Gain, M., Debnath, R., Adhikary, A., Qiao, Y., Hassan, M.M., Bairagi, A.K., Islam, S.M.S., 2024. Attention to monkeypox: An 41 inter- pretable monkeypox detection technique using attention mecha- nism. IEEE Access .https://doi.org/10.1109/ACCESS.2024.3385099

    2024

  15. [24]

    Monkeypox virus detection using pre- trained deep learning-based approaches

    Sitaula, C., Shahi, T.B., 2022. Monkeypox virus detection using pre- trained deep learning-based approaches. Journal of Medical Systems 46, 78.https://doi.org/10.1007/s10916-022-01868-2

    2022 doi

  16. [25]

    Monkeypox de- tection using deep neural networks

    Sorayaie Azar, A., Naemi, A., Babaei Rikan, S., Bagherzadeh Mohasefi, J., Pirnejad, H., Wiil, U.K., 2023. Monkeypox de- tection using deep neural networks. BMC Infectious Diseases 23, 438.https://doi.org/10.1186/s12879-023-08408-4

    2023 doi

  17. [26]

    Hyper-parameter tuned deep learning approach for ef- fective human monkeypox disease detection

    Dahiya, N., Sharma, Y.K., Rani, U., Hussain, S., Nabilal, K.V., Mohan, A., Nuristani, N., 2023. Hyper-parameter tuned deep learning approach for ef- fective human monkeypox disease detection . Scientific reports 13, 15930.https://doi.org/10.1038/s41598-023-43236-1

    2023 doi

  18. [28]

    Attention learning models using local zernike moments-based normalized images and convolutional neural networks for skin lesion classification

    Singh, C., Ranade, S.K., Singh, S.P., 2024. Attention learning models using local zernike moments-based normalized images and convolutional neural networks for skin lesion classification. Biomedical Signal Process- ing and Control 96, 106512.https://doi.org/10.1016/j.bspc.2024.106512

    2024

  19. [29]

    Mox-net: Multi-stage deep hybrid feature fusion and selection frame- work for mon- keypox classification

    Maqsood, S., Dama ˇseviˇcius, R., Shahid, S., Forkert, N.D., 2024. Mox-net: Multi-stage deep hybrid feature fusion and selection frame- work for mon- keypox classification. Expert Systems with Applications 255, 124584. https://doi.org/10.1016/J.ESW A.2024.124584

    2024

  20. [30]

    S., Hossain, M

    Bala, D., Hossain, M. S., Hossain, M. A., Abdullah, M. I., Rah- man, M. M., Manavalan, B., Gu, N., Islam, M. S., & Huang, Z. (2023). MonkeyNet: A robust deep convolutional neural network for monkeypox disease detection and classification. Neural Networks, 161, 757–775.https://...

    2023 doi

  21. [31]

    Meta- heuristics optimization-based ensemble of deep neural net- works for mpox disease de- tection

    Asif, S., Zhao, M., Tang, F., Zhu, Y., Zhao, B., 2023. Meta- heuristics optimization-based ensemble of deep neural net- works for mpox disease de- tection. Neural Networks 167, 342–359.https://doi.org/10.1016/j.neunet.2023.08.035 42

    2023 doi

  22. [32]

    Y.,& Sahin, V

    Oztel, I., Oztel, G. Y.,& Sahin, V. H. (2023). Deep Learning-Based Skin Diseases Classification using Smartphones. Advanced Intelligent Sys- tems, 5(12). https://doi.org/10.1002/aisy.202300211

    2023 doi

  23. [33]

    B.G., Rombach, H.D., 1994

    Caldiera, V.R. B.G., Rombach, H.D., 1994. The goal question metric approach. Encyclopedia of software engineering , 528–532

    1994

  24. [34]

    The effectiveness of data augmenta- tion in image classification using deep learning

    Wang, J., Perez, L., et al., 2017. The effectiveness of data augmenta- tion in image classification using deep learning. Convolutional Neural Networks Vis. Recognit 11, 1–8

    2017

  25. [35]

    Improving deep learning with generic data aug- mentation , in: 2018 IEEE sympo- sium series on computational intelligence (SSCI), IEEE

    Taylor, L., Nitschke, G., 2018. Improving deep learning with generic data aug- mentation , in: 2018 IEEE sympo- sium series on computational intelligence (SSCI), IEEE. pp. 1542–1545.https://doi.org/10.1109/SSCI.2018.8628742

    2018

  26. [37]

    Very deep convolutional networks for large-scale image recognition

    Simonyan, K., Zisserman, A., 2014. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556

    2014 arXiv

  27. [38]

    4510– 4520.https://doi.org/10.1109/CVPR.2018.00474

    Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., Chen, L.C., 2018.Mo- bilenetv2: Inverted residuals and linear bottlenecks , in: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 4510– 4520.https://doi.org/10.1109/CVPR.2018.00474

    2018

  28. [39]

    Re- thinking the inception architecture for computer vision , in: Proceedings of the IEEE conference on computer vision and pattern recognition, pp

    Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., Wojna, Z., 2016. Re- thinking the inception architecture for computer vision , in: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 2818–2826.https://doi.org/10.1109/CVPR.2016.308

    2016 doi

  29. [40]

    Ahsan, M.M., Li, Y., Zhang, J., Ahad, M.T., Yazdan, M.M.S.,

  30. [41]

    Optimization methods for large- scale machine learning

    Bottou, L., Curtis, F.E., Nocedal, J., 2018. Optimization methods for large- scale machine learning . SIAM review 60, 223–311

    2018

  31. [42]

    Adam: A method for stochastic optimiza- tion

    Kingma, D.P., Ba, J., 2015. Adam: A method for stochastic optimiza- tion. in- ternational conference on learning representations (2015) . San Diego, Cali- fornia .https://arxiv.org/abs/1412.6980v9

    2015 arXiv

  32. [43]

    R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Ba- tra, D

    Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Ba- tra, D. (2016).Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization.https://doi.org/10.1007/s11263-019-01228- 7

    2016 doi

  33. [44]

    Multimedia Tools and Applications 83, 71909– 71923

    Yolcu Oztel, G., 2024.Vision transformer and cnn-based skin lesion anal- ysis: classification of monkeypox. Multimedia Tools and Applications 83, 71909– 71923. https://doi.org/10.1007/s11042-024-19757-w

    2024 doi

  34. [45]

    The effectiveness of data augmentation in im- age classification using deep learning

    Perez, L., Wang, J., 2017. The effectiveness of data augmentation in im- age classification using deep learning. arXiv preprint arXiv:1712.04621

    2017 arXiv

  35. [46]

    External validity

    Findley, M.G., Kikuta, K., Denly, M., 2021. External validity. Annual review of political science 24, 365–393.https://doi.org/10.1146/annurev- polisci-041719-102556

    2021 doi

  36. [47]

    The precision-recall plot is more informative than the roc plot when evaluating bi- nary classifiers on imbalanced datasets

    Saito, T., Rehmsmeier, M., 2015. The precision-recall plot is more informative than the roc plot when evaluating bi- nary classifiers on imbalanced datasets. PloS one 10, e0118432. https://doi.org/10.1371/JOURNAL.PONE.0118432

    2015 doi

  37. [48]

    The class imbalance problem in deep learning

    Ghosh, K., Bellinger, C., Corizzo, R., Branco, P., Krawczyk, B., Jap- kowicz, N., 2024. The class imbalance problem in deep learning. Machine Learning 113, 4845–4901. https://doi.org/10.1007/S10994-022-06268- 8/FIGURES/27

    2024 doi

  38. [49]

    N., Ahmed, Md

    Ali, S. N., Ahmed, Md. T., Jahan, T., Paul, J., Sani, S. M. S., Noor, N., Asma, A. N., & Hasan, T. (2023). A Web-based Mpox Skin Lesion Detection System Using State-of-the-art Deep Learning Models Consid- ering Racial Diversity . https://arxiv.org/abs/2306.14169v1 44

    2023 arXiv

  39. [50]

    N., Ahmed, Md

    Ali, S. N., Ahmed, Md. T., Paul, J., Jahan, T., Sani, S. M. S., Noor, N., & Hasan, T. (2022).Monkeypox Skin Lesion Detection Using Deep Learn- ing Models: A Feasibility Study . https://arxiv.org/abs/2207.03342v1 45

    2022 arXiv

  40. [2020]

    Journal of Sensor and Actuator Networks 9, 54.https://doi.org/10.3390/jsan9040054 43

    Face recognition in an unconstrained and real-time envi- ronment using novel bmc- lbph methods incorporates with dji vision sensor . Journal of Sensor and Actuator Networks 9, 54.https://doi.org/10.3390/jsan9040054 43

    doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.