Enhancing Skin Disease Classification Through GAN Generate Synthetic images for improved CNN Training and Generalization

Every part of the world carries skin diseases which span from minor to potentially fatal conditions including skin cancer. Accurate early diagnosis of diseases depends on a time-intensive manual approach that shows errors frequently during the process. The proposed project develops an AI system to categorize skin conditions through deep learning using ResNet50 plus Random Forest as machine learning for image diagnosis alongside patient metadata examination. Explicitly the hybrid combined a CNN model to benchmark image classes while implementing Random Forest for metadata evaluation. The model benefits from enhanced generalization through GAN-generated synthetic images which help balance the dataset. The Streamlit-based web application enables users to upload images followed by diagnostic predictions output. The implemented system provides exceptional classification outcomes that allow medical staff to identify skin afflictions at the initial stage. The upcoming work plan consists of enlarging the present dataset while utilizing Explainable AI (XAI) methods and deploying the model on cloud services for simple access..