Varicella zoster early detection with deep learning

When Varicella-zoster virus affects the trigeminal nerve, early detection is important to prevent eye damage. Our study introduces a new methodology using convolutional neural networks, via a deep learning process, for the early detection of Herpes Zoster virus (HZV). The methodology was constructed with preprocessing, segmentation, extraction, and classification stages. The varicella-zoster virus (VZV) is responsible for causing two diseases: Varicella (also known as chickenpox) and Herpes Zoster (HZ). HZ is associated with the trigeminal nerve on one side of the face. In this paper are determined the limitations, devices, and software failures associated with developing early HZ disease detection system. A main factor is the classification process, which should enable the categorization of data training. Herein, we incorporated the techniques of K-Nearest Neighborhood (KNN), neural networks, and logistic regression for this purpose, the effectiveness of the paradigm on average was 97% for early detection with minimal information.