Monkeypox detection from skin lesion images using an amalgamation of CNN models aided with Beta function-based...

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Disease Research

Background:

• The COVID-19 pandemic highlighted the need for robust public health preparedness against infectious diseases.
• Monkeypox poses a significant global health threat, necessitating efficient diagnostic tools for early detection and management.

Purpose of the Study:

• To develop and evaluate an ensemble learning framework for accurate Monkeypox virus detection from skin lesion images.
• To improve early diagnosis capabilities for infectious disease outbreaks.

Main Methods:

• Utilized three pre-trained deep learning models (Inception V3, Xception, DenseNet169) fine-tuned on Monkeypox image data.
• Implemented an ensemble approach combining model predictions using Beta function-based probability normalization and a sum rule.
• Evaluated the framework on a public Monkeypox skin lesion dataset with a five-fold cross-validation.

Main Results:

• The ensemble model achieved an average accuracy of 93.39%.
• High performance metrics including precision (88.91%), recall (96.78%), and F1 score (92.35%) were recorded.
• The framework demonstrated effectiveness in distinguishing Monkeypox from other conditions based on skin lesions.

Conclusions:

• The proposed ensemble learning framework offers a promising tool for the early and accurate detection of Monkeypox.
• This AI-driven approach can enhance pandemic response strategies by enabling rapid diagnosis.
• Further research can explore integrating this framework into broader public health surveillance systems.