Lost And Found

Inspiration At universities, losing valuable items can be stressful, and finding their rightful owners is often difficult. We noticed a gap in communication and organization when it comes to returning lost items at UMT. This inspired us to build a Lost and Found Portal tailored for our campus—an efficient system to report, track, and claim items. With a touch of AI image detection, we aimed to make the matching process smarter and faster.

What it does Lost and Found is a full-stack web portal where UMT students and staff can:

Post lost or found items with descriptions, images, location, and contact info.

Browse or filter listings by category, date, or location.

Claim items through a request system that notifies the finder.

Receive admin approvals for moderation and prevent spam.

Use AI-based image detection to match similar lost/found item images for better accuracy.

Track requests and approvals from personalized dashboards.

How we built it Frontend: React.js with TailwindCSS and Ant Design for a responsive and modern UI.

Backend: Node.js with Express.js, MongoDB for the database, and Mongoose for data modeling.

Authentication: JWT-based login with role-based access (Student, Admin).

Image Handling: Multer for file uploads, integrated with Cloudinary for image storage.

AI Image Detection: Integrated a similarity comparison using TensorFlow.js and pre-trained models to identify visually similar items.

Real-time Chat & Notifications (optional): Enabled via Socket.io for communication between finders and claimers.

Challenges we ran into Implementing AI-based image comparison in a short time frame and limited compute resources.

Ensuring secure, role-based access and permissions.

Managing image uploads and previews efficiently.

Handling data filtering and dashboard UX for different user roles.

Integrating various tech stacks (AI + MERN) without breaking the flow.

Accomplishments that we're proud of Built a fully functional MERN app with clean UI/UX in a limited time.

Integrated AI image detection to bring innovation into a traditional lost and found system.

Implemented a working claim request flow with user notifications.

Created dashboards for both students and admins, handling real use cases.

Made something that could genuinely help our university community.

What we learned Hands-on experience with full-stack development, especially with user authentication and role-based dashboards.

Learned how to integrate AI/ML models into web apps for real-world use cases.

Understood how to manage cloud storage for image-heavy applications.

Improved teamwork, debugging, and feature prioritization under hackathon pressure.

What's next for Lost and Found Train and fine-tune our image matching model for better accuracy.

Integrate real-time chat between users for smoother communication.

Add location-based search using maps.

Deploy the platform publicly and collaborate with UMT administration for real-world adoption.

Introduce verification badges for trusted users to improve item return success rates.

Built With

• mern
• tailwind