-
-
logo
Inspiration
Lost and found systems have remained largely unchanged for decades, relying on manual descriptions and human memory to match items. We've all experienced the frustration of losing something valuable and the inefficiency of traditional lost and found processes. With the advancement of AI and computer vision technology, we saw an opportunity to revolutionize how people reconnect with their lost belongings using precise visual matching.
What it does
Lost and Found is a program designed to help users report and locate lost items efficiently. Using advanced video recognition technology, it identifies close matches and provides navigation assistance to guide users to their lost belongings.
How we built it
-Python for our frontend and backend, utilizing streamlit for the frontend and fastapi for our backend
-OpenCV was used in order to have a base model for our computer vision, which we then modified for our own purposes
-ResNet - used to determine object similarity
- YOLOv8 - used to identify individual objects in images
-MongoDB – Used for all things related to our lost and found items, except for storing the images of the objects themselves
-Azure Blob Storage & ACR – Used to store our lost item images
-Docker & Terraform – Utilized Docker, Terraform and many azure services to enable us to be able to deploy not only our backend but our frontend
-CI/CD & GitHub Actions – Automated deployment and continuous integration
Challenges we ran into
- Easily the greatest challenge was managing the complex UI interactions and states beyond typical use cases with Streamlit
-Handling API calls such as Google Maps and using these APIs with streamlit
-Azure was giving us many problems as well, especially when it came to deploying our backend
Accomplishments that we're proud of
We're proud of all the effort combined together to bring a robust application containing impressive features such as our AI object recognition and a good use of our technologies. We set so many goals along the way and managed to do way more than we could have imagined. It feels nice to reach the goals we set for ourselves.
What we learned
We learned that we should always keep an eye on our API credits and that learning a new framework is never easy (duh!)
What's next for 404 Lost & Found
We will keep improving our AI recognition model to keep finding lost objects as effectively as possible.
All in all, we are hoping that 404 will be able to save people from having bad days and losing some of their precious belonging. We will keep improving our application to keep making people's life better.
Log in or sign up for Devpost to join the conversation.