Acne Scan
- Processed 2500+ of labeled dermatology images, performing data augmentation and normalization to train a model with MobileNetV2, achieving ~45% accuracy in identifying multiple acne categories
- Developed a full-stack dermatology web application using Next.js/React, integrating Supabase for authentication, secure data storage, and real-time updates, enabling users to upload images and receive acne analysis.
TypescriptNext.jsReact.jsSupaBaseTensorFlowMobileNetV2Flask
View Demo