Monkeypox Detection Using Transfer Learning, ResNet50, Alex Net, ResNet18 & Custom CNN Model

Abstract

The latest monkeypox flare-up has arisen as a general well-being worry because of the fast spread to more than 40 countries that are not situated in Africa. Because of likenesses with chickenpox and measles, the early clinical ID of monkeypox can challenge. PC helped recognition of monkeypox sores might be valuable for checking and fast recognizable proof of thought situations when corroborative Polymerase Chain Reaction (PCR) tests are inaccessible. At the point when there are sufficient preparation models, profound learning strategies have been demonstrated to be helpful for naturally distinguishing skin injuries. Four pre-prepared Convolutional Neural Network (CNN) models are utilized to assess the exhibition of the transfer learning strategy: ResNet50, AlexNet,  and ResNet18 An unrivaled association considering the merging of ResNet18 and Google Net is similarly suggested. The recommended network can accomplish 91.57% exactness, 85.69% awareness, particularity and accuracy, individually. The proposed strategy outflanks every individual CNN for monkeypox identification.