Development of a Convolutional Neural Network-Based Web Application for Automated Skin Disease Classification
Dermatological disorders remain a significant global healthcare challenge, affecting millions of individuals and contributing to increased disease burden, particularly when delayed or inaccurate diagnosis affects treatment outcomes. Although artificial intelligence has demonstrated substantial potential in supporting dermatological diagnosis, many existing approaches are constrained by limited dataset diversity and insufficient deployment into practical healthcare environments. This study aimed to develop a Convolutional Neural Network (CNN) based web application for automated skin disease classification using dermoscopic images. A deep learning framework was implemented to automatically extract discriminative visual features from annotated dermoscopic skin lesion datasets. Image preprocessing techniques, including resizing, normalization, and data augmentation, were applied to enhance model generalization and reduce overfitting. The developed CNN model was evaluated using standard classification metrics, including accuracy, precision, recall, and F1-score, and subsequently integrated into a web-based application to enable real-time disease prediction. Experimental evaluation demonstrated that the system effectively classified multiple skin disease categories and provided automated predictions through an accessible web interface. The deployment of the trained model improved practical usability by providing a rapid and intelligent decision-support tool for preliminary dermatological assessment. The study highlights the potential of CNN based image classification and web-enabled artificial intelligence systems in supporting early skin disease detection, improving diagnostic accessibility, and enhancing clinical decision-support capabilities, particularly in healthcare settings with limited access to dermatology specialists.