Smart Plate

System Architecture Diagram

Our product is a web-based food waste reduction platform designed to encourage users to finish their meals and reduce unnecessary waste. The system works by awarding or deducting points based on how much food a user leaves behind. Those points can later be used to redeem prizes, turning sustainable behavior into something rewarding and engaging. The product has two main roles: users and restaurant owners. Users can log in, generate a QR code from their account, and present it at a food store when ordering. Restaurant owners log in through a separate interface, scan the user’s QR code, and assign a plate number and dish name to that user. This creates an order record in the backend so the plate can later be identified during tray return. After the user finishes eating, the restaurant owner or staff member can scan the returned tray using the camera interface on the website. The tray image is sent to the backend, where the Reka API is used to analyze the image and detect the plate number, waste level, waste percentage, and wasted food items. Based on this analysis, the system updates the user’s points, stores the scan result, and closes the corresponding order. We built the backend using Python and FastAPI, which provides clean and fast API endpoints for authentication, QR generation/decoding, order creation, and tray scan analysis. We used Supabase as our database and authentication layer to store user profiles, QR bindings, orders, and tray scan records. Supabase also handles authentication so users and restaurant owners can log in securely. The frontend communicates with the backend through HTTP API calls. This makes the system modular: the frontend focuses on the user interface and camera scanning, while the backend handles business logic, database access, and image analysis integration. We also used QR code generation and decoding services so the user QR flow works end-to-end. Overall, this product combines web development, backend engineering, database design, authentication, QR code workflows, and AI-powered image analysis to solve a real-world sustainability problem in a practical and scalable way.