Enhancing monkeypox diagnosis: A machine learning approach for skin lesion classification

Abstract

Monkeypox, a skin illness caused by the varicellazoster virus, is the focus of this research, which explores the potential of a Machine Learning (ML) system for classifying and detecting the disease. The dataset, sourced from Kaggle, consists of images depicting monkeypox lesions, which are augmented to develop and test custom models. Additionally, a web app is created, enabling users to submit images for analysis and classification by the machine learning models. The primary objective is to assess the utility and effectiveness of applying ML models for categorizing and identifying monkeypoxe. The study compares the performance of ResNet50, InceptionV3, Xception model, DenseNet121, and MobileNet, revealing improved accuracy, precision, recall, and F-1 score in the MobileNet and Xception models. The results present a confusion matrix, with MobileNet demonstrating a mean accuracy of 0.97, precision of 0.96, F-1 score of 0.968, and mean recall of 0.968.The main aim of this study is to assess the usefulness or effectiveness assessing the efficacy of utilizing machine learning models for the purpose of categorization and evaluation.