Deep Learning-Based Classification of Viral Dermatological Infections: Chickenpox, Measles, and Monkeypox

Skin infections, such as chickenpox, monkeypox, and measles, are of great importance to public health, particularly in developing and resource-constrained countries where access to early diagnosis and treatment is sometimes limited. The current COVID-19 pandemic has worsened the outbreak of monkeypox, underlining the critical requirement for accurate and timely diagnostic tools. Some misdiagnoses or delayed treatment may lead to serious health consequences in the case of certain diseases. This work proposes a deep learning-based classification model that would fulfill the need for effective diagnostic methodologies, especially in the early diagnosis of skin diseases and, more specifically, viral ones like measles, chickenpox, and monkeypox. It achieves high classification accuracy using convolutional neural networks for the MSID dataset, which consists of 770 photos of four classes: Measles, Normal, Chickenpox, and Monkeypox. In this paper, we overcome the strong constraints due to a tiny sample size using sophisticated data augmentation techniques that greatly improved our model's performance. The proposed system has very good performance, with validation accuracy of 98.59% and training accuracy of 99.90%. At last, this work contributes to the improvement of patient outcomes and the fostering of global public health activities by providing a trustworthy and effective diagnostic tool in support of global health efforts to eradicate infectious diseases.