👥 Princeton Hackathon Project – Facial Recognition System for Dementia Care
This project was developed as a feature for the Princeton Hackathon to assist caregivers in monitoring dementia patients by identifying whether a person is familiar or unfamiliar through live camera input.
The API uses DeepFace and OpenCV to perform real-time facial recognition (~97% accuracy). It powers a full-stack web app that alerts caregivers if an unknown person is detected or if the patient leaves a safe zone.
- Real-time facial recognition via webcam
- API returns "familiar" or "unfamiliar" status based on face matching
- Can be integrated with geofencing and caregiver alert systems
- Built for extensibility and deployment (Render-compatible)
- Backend: Flask, Python
- ML Library: DeepFace, OpenCV
- Deployment: Render
Built by Potri Abhisri Barama and team at Princeton Hackathon 2025.