Automated Monkeypox Skin Lesion Detection Using Deep Learning and Transfer Learning Techniques
Abstract
:1. Introduction
2. Related Work
3. Materials and Proposed Methods
3.1. Data Collection
3.2. Preprocessing
3.2.1. Image Resizing and Image Normalization
3.2.2. Data Augmentation
3.3. Model Selection and Training
3.3.1. Deep Convolutional Generative Adversarial Network (DCGAN)
Algorithm 1: Deep Learning Image Classification | |
//main steps in monkeypox image classification Input: Dataset Output: A network model: Model | |
1. | Begin |
//Data Collection: Data were gathered manually (45 images of monkeypox, 72 images of other diseases) | |
2. | Data <-- Normalization (Data) |
3. | Training, Validation, Test <-- Split Data (Data) |
//Data processing: Data augmentation (rotate, zoom in, zoom out, shift, share, and flip) using (TensorFlow. karas. preprocessing. Image) | |
Scale all images to (224 × 224) | |
4. | For i = 1 to Number of Training do |
5. | Data1 <-- Horizontal Flip (Training) |
6. | Data2 <-- Vertical Flip (Training) |
7. | Data3 <-- ZoomIn (Training) |
8. | Data4 <-- Zoom Out (Training) |
9. | Data5 <-- Shift (Training) |
10. | //Data Augmentation <-- Add (Training, Data1, Data2, Data3, Data4, Data5) |
11. | generator = make_generator_model () |
12. | generated_image = generator (noise, training) |
13. | discriminator = make_discriminator_model () |
14. | decision = discriminator (generated_image) |
//Training model: Divide the dataset into 80% for training and 20% for testing. Validation (chose 20% of images from training images) | |
15 | Model <-- Training Model (Data Augmentation, Validation) |
16. | Model <-- Training Evaluation (Model, Test) |
//Deep learning image classification (Determine if the photo contains monkeypox or not){EfficientNetB3, ResNet50, MobileNetV2, VGG16} | |
17. | End |
3.3.2. Pretrained DL Models
- I.
- VGG16 and VGG19
- II.
- ResNet
- III.
- MobileNet
- IV.
- EfficientNetB3
Input | Operator | T | c | n | s |
---|---|---|---|---|---|
2242 × 3 | Conv2d | - | 32 | 1 | 2 |
1122 × 32 | Bottleneck | 1 | 16 | 1 | 1 |
1122 × 16 | Bottleneck | 6 | 24 | 2 | 2 |
562 × 24 | Bottleneck | 6 | 32 | 3 | 2 |
282 × 32 | Bottleneck | 6 | 64 | 4 | 2 |
142 × 64 | Bottleneck | 6 | 96 | 3 | 1 |
142 × 96 | Bottleneck | 6 | 160 | 3 | 2 |
72 × 160 | Bottleneck | 6 | 320 | 1 | 1 |
72 × 320 | Conv2d 1 × 1 | - | 1280 | 1 | 1 |
72 × 1280 | Avgpool 7 × 7 | - | - | 1 | - |
1 × 1 × 1280 | Conv2d 1 × 1 | - | k | - |
3.4. Evaluation and Comparison
4. Results
Comparison of the Performance of DL Models with Our Models
5. Discussion
5.1. Validation of Our Model with Another Dataset
5.2. Comparison with Other Studies
6. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- Hussain, M.A.; Islam, T.; Chowdhury , F.U.H.; Islam, B.R. Can Artificial Intelligence Detect Monkeypox from Digital Skin Images? bioRxiv 2022. [Google Scholar] [CrossRef]
- Moore, M.J.; Rathish, B.; Zahra, F. Mpox (Monkeypox). In StatPearls; StatPearls Publishing: Tampa, FL, USA, 2022. [Google Scholar]
- Thornhill, J.P.; Barkati, S.; Walmsley, S.; Rockstroh, J.; Antinori, A.; Harrison, L.B.; Palich, P.; Nori, A.; Reeves, I.; Orkin, C.M.; et al. Monkeypox Virus Infection in Humans across 16 Countries—April–June 2022. N. Engl. J. Med. 2022, 387, 679–691. [Google Scholar] [CrossRef] [PubMed]
- Monkeypox. Available online: https://www.who.int/news-room/fact-sheets/detail/monkeypox (accessed on 23 January 2023).
- Gong, Q.; Wang, C.; Chuai, X.; Chiu, S. Monkeypox virus: A re-emergent threat to humans. Virol. Sin. 2022, 37, 477–482. [Google Scholar] [CrossRef] [PubMed]
- Zuluaga-Gomez, J.; Al Masry, Z.; Benaggoune, K.; Meraghni, S.; Zerhouni, N. A CNN-based methodology for breast cancer diagnosis using thermal images. Comput. Methods Biomech. Biomed. Eng. Imaging Vis. 2021, 9, 131–145. [Google Scholar] [CrossRef]
- Ali, S.N.; Ahmed, M.; Paul, J.; Jahan, T.; Sani, S.M.; Noor, N.; Hasan, T. Monkeypox Skin Lesion Detection Using Deep Learning Models: A Feasibility Study. arXiv 2022, arXiv:2207.03342. [Google Scholar]
- Sahin, V.H.; Oztel, I.; Oztel, G.Y. Human Monkeypox Classification from Skin Lesion Images with Deep Pre-trained Network using Mobile Application. J. Med. Syst. 2022, 46, 79. [Google Scholar] [CrossRef] [PubMed]
- Ahsan, M.M.; Uddin, M.; Farjana, M.; Sakib, A.; Al Momin, K.; Luna, S.A. Image Data Collection and Implementation of Deep Learning-Based Model in Detecting Monkeypox Disease Using Modified VGG16. arXiv 2022, arXiv:2206.01862. [Google Scholar]
- Sitaula, C.; Shahi, T.B. Monkeypox Virus Detection Using Pre-trained Deep Learning-based Approaches. J. Med. Syst. 2022, 46, 78. [Google Scholar] [CrossRef] [PubMed]
- Kumar, V. Analysis of CNN features with multiple machine learning classifiers in diagnosis of monkepox from digital skin images. medRxiv 2022. [Google Scholar] [CrossRef]
- Yasmin, F.; Hassan, M.; Zaman, S.; Aung, S.; Karim, A.; Azam, S. A Forecasting Prognosis of the Monkeypox Outbreak Based on a Comprehensive Statistical and Regression Analysis. Computation 2022, 10, 177. [Google Scholar] [CrossRef]
- Abdelhamid, A.A.; El-Kenawy, E.-S.M.; Khodadadi, N.; Mirjalili, S.; Khafaga, D.S.; Alharbi, A.H.; Ibrahim, A.; Eid, M.M.; Saber, M. Classification of Monkeypox Images Based on Transfer Learning and the Al-Biruni Earth Radius Optimization Algorithm. Mathematics 2022, 10, 3614. [Google Scholar] [CrossRef]
- Altun, M.; Gürüler, H.; Özkaraca, O.; Khan, F.; Khan, J. Monkeypox Detection Using CNN with Transfer Learning. Sensors 2023, 23, 1783. [Google Scholar] [CrossRef] [PubMed]
- Mohbey, K.K.; Meena, G.; Kumar, S.; Lokesh, K. A CNN-LSTM-based hybrid deep learning approach to detect sentiment polarities on Monkeypox tweets. arXiv 2022. [Google Scholar] [CrossRef]
- Esteva, A.; Kuprel, B.; Novoa, R.A.; Ko, J.; Swetter, S.M.; Blau, H.M.; Thrun, S. Dermatologist-level classification of skin cancer with deep neural networks. Nature 2017, 542, 115–118. [Google Scholar] [CrossRef] [PubMed]
- Data_Monkeypox|Kaggle. Available online: https://www.kaggle.com/datasets/ahmadnasayrah/data-monkeypox (accessed on 23 January 2023).
- Ying, X. An Overview of Overfitting and its Solutions. J. Phys. Conf. Ser. 2019, 1168, 022022. [Google Scholar] [CrossRef]
- Simonyan, K.; Zisserman, A. Very deep convolutional networks for large-scale image recognition. In Proceedings of the 3rd International Conference on Learning Representations, ICLR 2015—Conference Track Proceedings, San Diego, CA, USA, 7–9 May 2015; pp. 1–14. [Google Scholar]
- He, K.; Zhang, X.; Ren, S.; Sun, J. Deep residual learning for image recognition. In Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 27–30 June 2016; Volume 2016, pp. 770–778. [Google Scholar] [CrossRef] [Green Version]
- Tan, M.; Le, Q.V. EfficientNet: Rethinking model scaling for convolutional neural networks. In Proceedings of the 36th International Conference on Machine Learning, PMLR, Long Beach, CA, USA, 10–15 June 2019; Volume 2019, pp. 10691–10700. [Google Scholar]
- Sandler, M.; Howard, A.; Zhu, M.; Zhmoginov, A.; Chen, L.C. MobileNetV2: Inverted Residuals and Linear Bottlenecks. In Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–23 June 2018; pp. 4510–4520. [Google Scholar] [CrossRef] [Green Version]
Author | Approaches | Best Results | Validation | Limitation and Gaps |
---|---|---|---|---|
[7] | VGG-16, ResNet50, and InceptionV3 | Accuracy: 82.96% | 70% training 10% validation 20% testing | Low accuracy and it can improve |
[8] | ResNet18, GoogleNet, EfficientNetB0, NasnetMobile, ShufeNet, and MobileNetV2 | Accuracy: 91.11%. | 70% training, 20% testing. 10% validation | Low accuracy and small dataset |
[9] | VGG16 | AUC: 97.2% | 80% training 20% testing | They did not use different models |
[10] | VGG-16, VGG-19, ResNet50, ResNet101, IncepResNetv2, MobileNetV2, InceptionV3, Xception, EfficientNetB0, EfficientNetB1, EfficientNetB2, DenseNet-121, and DenseNet-169 | Avg. Accuracy: 85.44% | No Validation | Low accuracy |
[11] | CNN models AlexNet, GoogleNet and Vgg16Net with, Naïve Bayes, SVM, KNN, Random Forest (RF), and Decision Tree (DT) | Accuracy: 91.11% | No Validation | No graph to confirm the results |
[12] | Polynomial Regression, SVR, Holt’s Linear Model AR Model, SARIMA Model ARIMA Model, MA Model, Holt–Winter’s Model, and Prophet Model | MSE: 41,922.55 R2: 0.49 MAPE: 16.82 MAE: 146.29 RMSE: 204.75 | No Validation | Low accuracy |
[13] | Binary PSOBER algorithm | Accuracy: 98.8% | No Validation | No limitation |
[14] | CNN model based on MobileNetV3-s, EfficientNetV2, ResNet50, Vgg19, DenseNet121, and Xception models. | Accuracy: 96%. | No Validation | No limitation |
[15] | CNN and LSTM | Accuracy: 94%. | 70% training, 30% validation | No limitation |
Type of Image | Original Images | Augmented Images |
---|---|---|
Monkeypox | 45 | 540 |
Normal | 25 | 300 |
Scarlet fever | 22 | 264 |
Roseola | 25 | 300 |
Prediction | |||
Positive: Monkeypox | Negative: No-Monkeypox | ||
Actual | Positive: Monkeypox | Number of TP | Number of FN |
Negative: No-Monkeypox | Number of FP | Number of TN |
Classifier | Training Time (Second) | Testing Time (Second) | Epoch Number | Best Epoch | Per Epoch | Best Training Accuracy | Best Validation Accuracy |
---|---|---|---|---|---|---|---|
EfficientNetB3 | 264 | 5 | 14 | 13 | 11 | 0.652 | 0.667 |
VGG19 | 297 | 9 | 14 | 12 | 11 | 0.984 | 0.934 |
VGG16 | 304 | 9 | 14 | 11 | 11 | 0.999 | 0.981 |
ResNet50 | 269 | 5 | 14 | 14 | 11 | 0.918 | 0.733 |
MobileNetV2 | 256 | 4 | 14 | 9 | 11 | 0.998 | 0.992 |
Classifier Model | Accuracy | Precision | Recall | F1-Score |
---|---|---|---|---|
EfficientNetB3 | 0.6838 | 0.38 | 0.72 | 0.49 |
VGG19 | 0.9779 | 0.96 | 0.98 | 0.97 |
VGG16 | 0.9669 | 0.95 | 0.96 | 0.95 |
ResNet50 | 0.7069 | 0.34 | 0.90 | 0.49 |
MobileNetV2 | 0.9816 | 0.99 | 0.96 | 0.98 |
Classifier of Models | Accuracy | Precision | Recall | F1-Score |
---|---|---|---|---|
EfficientNetB3 | 0.63 | 0.94 | 0.19 | 0.31 |
VGG19 | 0.91 | 0.94 | 0.85 | 0.89 |
VGG16 | 0.89 | 0.93 | 0.82 | 0.87 |
ResNet50 | 0.78 | 0.85 | 0.63 | 0.72 |
MobileNetV2 | 0.94 | 0.96 | 0.92 | 0.94 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Jaradat, A.S.; Al Mamlook, R.E.; Almakayeel, N.; Alharbe, N.; Almuflih, A.S.; Nasayreh, A.; Gharaibeh, H.; Gharaibeh, M.; Gharaibeh, A.; Bzizi, H. Automated Monkeypox Skin Lesion Detection Using Deep Learning and Transfer Learning Techniques. Int. J. Environ. Res. Public Health 2023, 20, 4422. https://doi.org/10.3390/ijerph20054422
Jaradat AS, Al Mamlook RE, Almakayeel N, Alharbe N, Almuflih AS, Nasayreh A, Gharaibeh H, Gharaibeh M, Gharaibeh A, Bzizi H. Automated Monkeypox Skin Lesion Detection Using Deep Learning and Transfer Learning Techniques. International Journal of Environmental Research and Public Health. 2023; 20(5):4422. https://doi.org/10.3390/ijerph20054422
Chicago/Turabian StyleJaradat, Ameera S., Rabia Emhamed Al Mamlook, Naif Almakayeel, Nawaf Alharbe, Ali Saeed Almuflih, Ahmad Nasayreh, Hasan Gharaibeh, Mohammad Gharaibeh, Ali Gharaibeh, and Hanin Bzizi. 2023. "Automated Monkeypox Skin Lesion Detection Using Deep Learning and Transfer Learning Techniques" International Journal of Environmental Research and Public Health 20, no. 5: 4422. https://doi.org/10.3390/ijerph20054422