Monkeypox Detection using CNN-Based Pretrained Models

Monkeypox is a zoonotic pathogen that causes a disease in humans and other mammals. The virus has spread rapidly in more than 113 countries as of 18 March 2023, sparking concerns about global healthcare. Early diagnosis is crucial to slow the disease's quick course. Computer-aided identification of monkeypox from skin lesion images may be useful in this scenario. In this work, the use of deep learning approaches is explored to helps in the early-stage recognition of the monkeypox virus. Various pre-trained models, including EfficientNetB3, ResNet50V2, VGG16, DenseNet121, and InceptionV3 are applied on the Monkeypox Skin Lesion Dataset (MSLD). The use of pre-trained models in the framework reduces the need for large amounts of training data and specialized computer resources, making it more accessible to healthcare providers in resource-limited settings and the proposed framework represents a potential strategy for enhancing the early detection and management of monkeypox. The EfficientNetB3 model achieved the highest accuracy of 98%. These findings have significant implications for the creation of automated methods for the early identification of monkeypox, which can enhance global healthcare.