Saint Peter’s Lost & Found

AI-powered Lost & Found that uses computer vision and text generation to detect, describe, and list campus items, making it easy for students to find, claim, or report lost belongings.

Comment

Inspiration

Every day, the campus inbox fills with emails about lost and found items (phones, keys, IDs, and more.) We wanted to simplify that process by creating one centralized platform where students and faculty can easily post, find, and claim lost items without all the back-and-forth emails.

What it does

Saint Peter’s Lost & Found is an AI-powered web platform that helps students and faculty easily find and report lost items on campus. Users can upload photos of found objects, and the system automatically detects what’s in the image using YOLO computer vision, then generates a title and description using a text-generation model. All items are stored in a searchable database where users can browse, claim, or report them. This will eliminate the flood of lost-item emails.

How we built it

We built the frontend using HTML and CSS to create a simple, responsive, and user-friendly interface for browsing and submitting items. The backend was developed with FastAPI, which connects to a YOLOv8 model for object detection. A text-generation model produces concise titles and descriptions for each detected item, and all data including photos, metadata, and detection confidence is stored in a structured database.

Challenges we ran into

Integrating computer vision and text generation smoothly in the backend.

Handling image uploads and data storage efficiently through FastAPI.

Ensuring the frontend communicated properly with the backend endpoints.

Balancing detection accuracy with performance for real-time uploads.

Accomplishments that we're proud of

Completed a working frontend in HTML/CSS with functional upload and browsing features.

Successfully integrated YOLO object detection and AI-generated descriptions into the workflow.

Built a fully connected backend and frontend pipeline for automated lost-and-found item management.

Created a foundation that can significantly reduce email clutter and streamline campus communications.

What we learned

How to integrate multiple AI models (vision + language) into one cohesive application.

Hands-on experience with FastAPI, YOLO, and frontend-backend communication.

The importance of user-friendly design in driving real-world adoption.

How automation can improve everyday campus operations.

What's next for Saint Peter’s Lost & Found

Add search and filter features by category or location.

Implement user accounts and notifications for claim tracking.

Improve database design for scalability and better item management.

Deploy the full system on a university server for real-world use by students and staff.

Built With

Share this project:

Updates