Vitiligo Detection Using Convolutional Neural Networks

Abstract—Vitiligo is a chronic dermatological condition characterized by depigmentation of the skin. This study proposes a Convolutional Neural Network (CNN)-based system for automated detection of vitiligo from skin images. The model uses a pipeline consisting of image preprocessing (normalization, resizing, augmentation), followed by a deep learning classifier trained to distinguish between healthy and vitiligo affected skin. The system is integrated into a three-tier web application (VitiScan) using React for frontend, Flask for backend, and a PyTorch-based model. Evaluation metrics including accuracy, precision, recall, and F1-score confirm the system's reliability. The proposed model enhances diagnostic capabilities and offers a potential assistive tool for dermatologists. Keywords—Vitiligo, Convolutional Neural Networks, Image classification, Deep Learning.