Project Status: Completed — Built unified deep learning pipelines for automated monkeypox detection, handling both binary classification (Mpox vs. others) (Nayak et al., 2023) and multiclass (Mpox, Chickenpox, Measles, Healthy) (Nayak et al., 2023) tasks from skin lesion images. Leveraged curated Kaggle datasets, enriching them with extensive augmentation methods for better model robustness. Used transfer learning to deploy ResNet-18, ResNet-50, ResNet-101, and SqueezeNet architectures in MATLAB, experimenting on consumer GPUs with careful tuning of mini-batch size and learning rate to maximize accuracy.
.
For binary detection, models hit up to 99.5% accuracy and F1 score, and for multiclass labeling, reached 91% accuracy and a 92.6% F1 on Mpox. The approach included explainable AI (LIME, GradCAM), letting clinicians see the reasoning behind predictions, not just black-box outputs. These solutions were coded and optimized to run efficiently on regular laptops and phones, so they fit real-world, potentially low-resource contexts where quick diagnostics are needed and PCR isn’t available. Extensive benchmarking showed these approaches are competitive—with accuracy and speed to rival or outperform existing literature—while still being practical and interpretable for the healthcare setting.
References
2023
-
Deep learning based detection of monkeypox virus using skin lesion images
Tushar Nayak, Krishnaraj Chadaga, Niranjana Sampathila, and 5 more authors
Medicine in Novel Technology and Devices, Jun 2023
As we set into the second half of 2022, the world is still recovering from the two-year COVID-19 pandemic. However, over the past three months, the outbreak of the Monkeypox Virus (MPV) has led to fifty-two thousand confirmed cases and over one hundred deaths. This caused the World Health Organisation to declare the outbreak a Public Health Emergency of International Concern (PHEIC). If this outbreak worsens, we could be looking at the Monkeypox virus causing the next global pandemic. As Monkeypox affects the human skin, the symptoms can be captured with regular imaging. Large samples of these images can be used as a training dataset for machine learning-based detection tools. Using a regular camera to capture the skin image of the infected person and running it against computer vision models is beneficial. In this research, we use deep learning to diagnose monkeypox from skin lesion images. Using a publicly available dataset, we tested the dataset on five pre-trained deep neural networks: GoogLeNet, Places365-GoogLeNet, SqueezeNet, AlexNet and ResNet-18. Hyperparameter was done to choose the best parameters. Performance metrics such as accuracy, precision, recall, f1-score and AUC were considered. Among the above models, ResNet18 was able to obtain the highest accuracy of 99.49%. The modified models obtained validation accuracies above 95%. The results prove that deep learning models such as the proposed model based on ResNet-18 can be deployed and can be crucial in battling the monkeypox virus. Since the used networks are optimized for efficiency, they can be used on performance limited devices such as smartphones with cameras. The addition of explainable artificial intelligence techniques LIME and GradCAM enables visual interpretation of the prediction made, helping health professionals using the model.
-
Detection of Monkeypox from skin lesion images using deep learning networks and explainable artificial intelligence
Tushar Nayak, Krishnaraj Chadaga, Niranjana Sampathila, and 5 more authors
Applied Mathematics in Science and Engineering, Jun 2023
Monkeypox (Mpox) resurfaced in January 2022 as a rare zoonotic disease that spreads to many countries. Though the virus is not as dangerous as COVID-19, it has still caused many fatalities worldwide. The Mpox virus spreads when people are in close contact with infected individuals. Among many symptoms, the disease also causes skin rashes, and medical imaging can be used to diagnose the virus successfully. However, other diseases such as smallpox, chickenpox, and measles also cause similar skin rashes. Hence, artificial intelligence (AI) and machine learning (ML) can be highly beneficial in diagnosing Mpox from other similar diseases. After extensive model training, it is advantageous to use a standard camera to capture skin images of an infected patient and run it against deep learning (DL) models. In this research, we have used transfer learning models such as residual networks and SqueezeNet to diagnose Mpox from measles, chickenpox and healthy patients. An average accuracy of 91.19% and an F1-score of 92.55% were obtained for the Mpox class. The findings show that the models can be useful in detecting the contagious virus. Since the classifiers are easily deployable, they can be used on camera-ready devices such as phones and laptops.
