Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Dec 7:14:1225557.
doi: 10.3389/fimmu.2023.1225557. eCollection 2023.

A study on the recognition of monkeypox infection based on deep convolutional neural networks

Junkang Chen  1 , Junying Han  1
Affiliations

A study on the recognition of monkeypox infection based on deep convolutional neural networks

Junkang Chen et al. Front Immunol. .

Abstract

Introduction: The World Health Organization (WHO) has assessed the global public risk of monkeypox as moderate, and 71 WHO member countries have reported more than 14,000 cases of monkeypox infection. At present, the identification of clinical symptoms of monkeypox mainly depends on traditional medical means, which has the problems of low detection efficiency and high detection cost. The deep learning algorithm is excellent in image recognition and can extract and recognize image features quickly and reliably.

Methods: Therefore, this paper proposes a residual convolutional neural network based on the λ function and contextual transformer (LaCTResNet) for the image recognition of monkeypox cases.

Results: The average recognition accuracy of the neural network model is 91.85%, which is 15.82% higher than that of the baseline model ResNet50 and better than the classical convolutional neural networks models such as AlexNet, VGG16, Inception-V3, and EfficientNet-B5.

Discussion: This method realizes high-precision identification of skin symptoms of the monkeypox virus to provide a fast and reliable auxiliary diagnosis method for monkeypox cases for front-line medical staff.

Keywords: aided diagnosis; contextual transformer; deep learning; monkeypox images; residual convolutional networks.

PubMed Disclaimer

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

Figure 1
Figure 1
Sample images of the dataset. (A) Varicella, (B) Cowpox, (C) Healthy, (D) Measles, (E) Monkeypox, (F) Variola.
Figure 2
Figure 2
The design idea of the backbone.
Figure 3
Figure 3
The overall architecture of the LaCTResNet model. The cube represents the output tensor; H, W, and Channel represent the height, width, and number of channels of the tensor, respectively. The circle represents the output category.
Figure 4
Figure 4
Residual structure.
Figure 5
Figure 5
Calculation diagram of the λ function layer.
Figure 6
Figure 6
(A) Calculation diagram of traditional attention mechanism, (B) Calculation diagram of contextual transformer layer.
Figure 7
Figure 7
The recognition accuracy and parameter quantity of the models. The vertical coordinate indicates the average recognition accuracy of the model, and the horizontal coordinate indicates the number of parameters of the model.
Figure 8
Figure 8
The recognition accuracy and loss curve of models on the training set. The vertical coordinates in (A) indicate the accuracy rate; the vertical coordinates in (B) indicate the loss value; the horizontal coordinates all indicate the training period.
Figure 9
Figure 9
Confusion matrix heat map. The number represents the total number of sample images predicted as that class by the model; the larger the value, the darker the color.
Figure 10
Figure 10
F1 Score of the LaCTResNet family of models.
Figure 11
Figure 11
ROC curve of LaCTResNet56 model.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Kumar N, Acharya A, Gendelman HE, Byrareddy SN. The 2022 outbreak and the pathobiology of the monkeypox virus. J Autoimmun (2022) 131:102855. doi: 10.1016/j.jaut.2022.102855 - DOI - PMC - PubMed
    1. Wang M, Cui F. Status of epidemiology and prevention and control of monkeypox. Jiangsu J Prev Med (2023) 34(01):8–11. doi: 10.13668/j.issn.1006-9070.2023.01.002 - DOI
    1. Shchelkunova GA, Shchelkunov SN. Smallpox, monkeypox and other human orthopoxvirus infections. Viruses (2022) 15(1):103. doi: 10.3390/v15010103 - DOI - PMC - PubMed
    1. Kumar P, Chaudhary B, Yadav N, Devi S, Pareek A, Alla S, et al. . Recent advances in research and management of human monkeypox virus: an emerging global health threat. (2023) 15(4):937. doi: 10.3390/v15040937 - DOI - PMC - PubMed
    1. Algarate S, Bueno J, Crusells MJ, Ara M, Alonso H, Alvarado E, et al. . Usefulness of non-skin samples in the PCR diagnosis of Mpox (Monkeypox)[J]. Viruses (2023) 15(5):1107. doi: 10.3390/v15051107 - DOI - PMC - PubMed

Publication types