Habiba Hye | 2025 I.S. Symposium

Name: Habiba Hye
Title: Scan, Click, Diagnose: An AI-Powered System for Real-Time Skin Condition Detection
Major: Computer Science
Minor: Studio Arts
Advisor: Heather Guarnera

My Independent Study explores the use of artificial intelligence to improve skin condition diagnosis through a real-time, web-based tool. By combining a deep learning model trained on medical-grade skin images (HAM10000, ISIC) with a machine learning model based on patient metadata (from the SCIN dataset), I created an ensemble system that achieved 91% accuracy—outperforming either model alone. What excites me most is how this project blends computer science and healthcare to solve real-world problems, especially in underserved communities. I learned that preprocessing and data quality are critical to model success, and I see future opportunities to expand the system with more diverse, real-world data and add explainability features to build user trust in AI-driven diagnostics.

Posted in Symposium 2025 on May 1, 2025.