A Comprehensive framework for classifying skin lesion diseases with class imbalance handling

Predictive diagnosis and classification of skin lesions are critical in managing viral infectious skin diseases like Measles, Monkeypox and Chickenpox. Traditional diagnostic methods, such as visual examination, histopathology, and laboratory tests, have been less effective in advancing medical field, and time consuming, also requires specialized expertise. This study explores the development of Deep Learning (DL) model to improve accuracy, efficiency and predictability of disease classification. By training on a large dataset of skin lesion images, a model has been trained on Convolutional Neural Network (CNN) with pretrained MobileNetV2 to predict and classify the disease. The model’s performance was evaluated against standard traditional methods, demonstrating its potential to provide scalable and more accurate diagnostic tools. This research highlights the need for improvising and moving to advanced classification methods, offering a comprehensive approach to skin disease classification and improving public health standards.