MRpoxNet: An enhanced deep learning approach for early detection of monkeypox using modified ResNet50.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Dermatology

Background:

• Early detection of monkeypox is crucial for effective public health response.
• Digital imaging offers a non-invasive method for diagnosing skin conditions.
• Deep learning models show promise in analyzing medical images for disease identification.

Purpose of the Study:

• To develop an enhanced deep learning model, MRpoxNet, for early monkeypox detection.
• To achieve high diagnostic accuracy and clinical reliability in identifying monkeypox from skin lesions.
• To leverage a modified ResNet50 architecture for improved performance.

Main Methods:

• Utilized an augmented dataset of 6116 skin lesion images (monkeypox, non-monkeypox, normal skin).
• Developed MRpoxNet by extending the ResNet50 architecture with additional layers.
• Evaluated performance using accuracy, precision, recall, F1 score, sensitivity, and specificity, comparing against established models.

Main Results:

• MRpoxNet achieved a diagnostic accuracy of 98.1%, surpassing baseline models.
• The model demonstrated superior robustness in distinguishing monkeypox lesions.
• All key performance metrics indicated high diagnostic reliability.

Conclusions:

• MRpoxNet offers a robust and efficient solution for early monkeypox detection.
• The model's performance suggests its suitability for integration into clinical diagnostic workflows.
• Future work includes dataset expansion and multimodal adaptability for diverse clinical scenarios.