Classification of monkeypox images using Al-Biruni earth radius optimization with deep convolutional neural network

There is a connection that has been established between the virus responsible for monkeypox and the formation of skin lesions. This virus has been detected in Africa for many years. Our research is centered around the detection of skin lesions as potential indicators of monkeypox during a pandemic. Our primary objective is to utilize metaheuristic optimization techniques to improve the performance of feature selection and classification algorithms. In order to accomplish this goal, we make use of deep learning and a transfer learning technique to extract attributes. The GoogleNet network, a deep learning framework, is used to carry out feature extraction. Furthermore, the feature selection process is conducted using a binary version of the dynamic Al-Biruni earth radius optimization (DBER). After that, the convolutional neural network is used to assign labels to the selected features from the collection. To improve the classification accuracy, adjustments are made to the convolutional neural network by utilizing the continuous version of the DBER algorithm. We used a range of metrics to analyze the different assessment methods, including accuracy, sensitivity, specificity, positive predictive value (P-value), negative predictive value (N-value), and F1-score. They were compared to each other. All the metrics, including the F1-score, sensitivity, specificity, P-value, and N-value, achieved high values of 0.992, 0.991, and 0.993, respectively. The outcomes were achieved by combining feature selection with the use of a convolutional neural network. After optimizing the parameters in the convolutional neural network, the proposed method achieved an impressive overall accuracy rate of 0.992.