Efficient Monkeypox Detection in Skin Lesions Using Pre-Trained Deep Learning Technique

A medical emergency was declared by the World Health Organization and the Department of Health when the monkeypox virus gained prominence during the COVID-19 pandemic. Traditional classification systems struggle to accurately identify skin disorders because viruses that cause them often have similar skin appearances. Concerns about security and privacy are often raised when sharing sensitive medical data. A workable solution to the issues related to medical data categorization and illness detection is provided by the combination of Federated Learning and deep neural networks. The proposed technique recommends using pre-trained Deep-Learning technique in a federated learning technique to identify the virus that infects human skin. The Kaggle and open-source datasets were both used. In order to develop a global technique, the research sought to collaborate with various local datasets while maintaining privacy concerns. The proposed technique is simultaneously computing the categorical accuracy level, F1-score micro, and F1-score macro. The effectiveness of the proposed technique is validated by comparing the procured results with the existing state-of-art techniques.