Inspiration
Shared fridges often lead to lost food, expired items, and confusion over ownership. We wanted a simple, low-cost way to add accountability and organization to shared spaces using computer vision.
What it does
Smart-Fridge is a vision-powered kiosk that tracks food deposits and retrievals. Users label items with generated codes, and the system verifies ownership using image fingerprinting. Suspicious removals trigger alerts, helping keep shared fridges fair and organized.
How we built it
We built Smart-Fridge in Python using OpenCV for item detection and fingerprinting, Tkinter for the kiosk UI, and SQLite for storage. A background subtraction pipeline isolates items, and a state machine manages door sessions and verification flows.
Challenges we ran into
Lighting variation and hand occlusion made reliable item detection tricky. We also had to design safeguards to prevent users from bypassing verification during a fridge session.
Accomplishments that we're proud of
We created a working end-to-end prototype that combines computer vision, UI design, and event tracking into a practical real-world system.
What we learned
We learned how to build robust computer vision pipelines and design user flows that handle real-world edge cases.
What's next for Smart-Fridge
We plan to deploy the system on ESP32 hardware, add expiry tracking and notifications, and improve recognition with lightweight ML models.
Log in or sign up for Devpost to join the conversation.