Human Pathogenic Monkeypox Disease Recognition Using Q-Learning Approach.

Area of Science:

• Artificial Intelligence
• Medical Imaging
• Machine Learning

Background:

• The monkeypox virus presents a growing global health concern, necessitating advanced diagnostic tools.
• Traditional methods for disease detection may not be sufficient for rapid and accurate monkeypox identification.
• Artificial intelligence (AI) offers potential solutions for enhancing the precision of medical image classification.

Purpose of the Study:

• To propose and evaluate two AI-driven strategies for improving monkeypox image classification accuracy.
• To leverage reinforcement learning and neural network parameter optimization for enhanced diagnostic capabilities.
• To provide tools for clinicians and public health agencies for monkeypox patient management and disease surveillance.

Main Methods:

• Development of two AI strategies focusing on feature extraction and classification for monkeypox detection.
• Implementation of the Q-learning algorithm for state-action rate determination and Malneural networks for parameter optimization.
• Evaluation of algorithms using an open-access dataset and interpretation criteria for feature selection analysis.

Main Results:

• Achieved high performance metrics for monkeypox classification: 95% precision, 95% recall, and 96% F1-score.
• The Malneural network demonstrated superior accuracy (approx. 0.985) compared to benchmark algorithms (DDQN, Policy Gradient, Actor-Critic).
• Proposed methods significantly outperformed traditional machine learning approaches in accuracy and effectiveness.

Conclusions:

• The proposed AI-based methods offer a significant advancement in monkeypox image classification accuracy.
• These techniques provide a robust framework for early and precise detection of monkeypox disease.
• The developed algorithms can support clinical decision-making and public health surveillance of monkeypox outbreaks.