A Framework of Deep Learning for Monkeypox Image Classification Based on Particle Swarm Optimization (PSO)

Abstract

Decisions made now will shape the future in which future generations will live. With all of this, people must be aware of the danger that faces every individual living on this earth. Among the dangers to which humanity is exposed are the spread of diseases, natural disasters, wars, and the like. Recently, many diseases have begun to spread, and on top of these diseases are Covid 19, Monkeypox (MPX), and many other diseases. This paper attempts to help many clinicians classify monkeypox by using different models in Artificial Intelligence (AI). Deep learning (DL) as a branch of artificial intelligence has been used as a proposed model to solve this problem. Among the elements used in the proposed model are data augmentation and extraction of various features by means of some VGG16 and Inception models. The Particle Swarm Optimization Algorithm (PSO)was also used to feature selection and optimization of the parameters of the neural network, while the classification process was carried out by Support Vector Machine (SVM). In the final process, the model had to be evaluated by a confusion matrix, which showed that the accuracy of the VGG16 model was 85% but after improved PSO, new accuracy become 94.5%. In the state of Inception model accuracy is 75.2% but when improved PSO is used it becomes 90.2%. These results are useful for the classification and diagnosis of monkeypox. It is no secret that personal hygiene, avoiding touching animal waste, and taking a vaccine to increase immunity are important ways to protect people from diseases.