Deep learning approach for monkeypox virus prediction: leveraging DensetNet-121 and image data

Abstract

The Mpox virus, sometimes referred to as monkeypox, causes flu-like symptoms and rashes. The variola virus, which causes smallpox, is linked to the virus that causes monkeypox. Smallpox symptoms are more severe than those of Mpox, and the illness is rarely deadly. There is no connection between Mpox and chickenpox. The variola virus of smallpox and the vaccinia virus being used in the smallpox vaccine both belong to the Orthopoxvirus genus, which also includes the uncommon viral disease known as monkeypox. This study aims to increase the effectiveness of monkeypox virus (MPV) identification by utilizing global historical records. This study examines several approaches and determines which produces the best results for the input data. Performance metrics have been used to compare the efficiency to current models. The underlying patterns and correlations in the data are then taught to Dense-Net-121 through the use of the training set. The remarkable results are as follows: accuracy at 96.12%, precision at 93.2%, recall at 90%, F1-score at 91%, the area under the curve-receiver operating characteristic (AUC-ROC) at 94.5%, and specificity at 94%, outperforming the existing methods.