Inspiration
The project was inspired by the challenge of managing lost belongings on the vast UMass Amherst campus and the need for a smarter alternative to traditional search methods.
What it does
Back2U utilizes CLIP-based semantic search to allow users to find lost items using natural language descriptions rather than exact keywords.
How we built it
We integrated OpenAI’s CLIP model with a React/Tailwind frontend and a Python/SQLite backend to map images and text into a shared vector space.
Challenges we ran into
Our primary hurdles were optimizing dual-perspective image encoding and refining similarity thresholds to ensure mathematically accurate search results.
Accomplishments that we're proud of
We are proud of creating a seamless vector-based matching system that successfully translates complex AI processes into an intuitive user experience.
What we learned
We mastered the implementation of high-dimensional embeddings and learned how to build robust, AI-integrated full-stack applications.
What's next for Back2U
Adding user accounts and claim flow.
Log in or sign up for Devpost to join the conversation.