THIS PROJECT DID NOT REACH FULL COMPLETION, AND IS UNOPERATIONAL AS WE COULD NOT CONNECT ALL THE PIECES IN TIME
Inspiration
We wanted to test the limits of AI visual understanding while creating a tool that can be amusing and useful.
What it does
We couldn't get the project fully working, but it was supposed to let the user take a photo, and RSU Zoner tells you whether you are in a Rural, Suburban, or Urban environment.
How we built it
We started by gathering 200 images for each zone type. Next we formatted the data to be more understandable and usable with a simple network we recycled from some old code. All the code mentioned so far was Python. Simultaneously, we were building an IOS app using Swift to be the front-end and acquire the users images. Lastly we compiled the python model to be usable on IOS and attempted to put it all the pieces together.
Challenges we ran into
The first challenge was of course data. Our dataset is extremely limited with only 2 people in a 36 hour window. The next large challenge was getting our network to work on IOS. Our model input format was tough to match on the IOS side, and that ended up being the bottle neck that stopped us from completion.
Accomplishments that we're proud of
We are proud to have come up with an idea, and pushed it through a development process with only 2 people. Our physical results are limited, but the idea and foundation are laid to allow further inspection upon retrieval of more data and some extra debugging.
What we learned
In the course of development errors can come out of nowhere that slow you down. Having an additional teammate or 2 would have likely helped.
What's next for RSU Zoner
Get more data. Fix the code. See if RSU Zoner can determine the different Rural, Suburban, and Urban zones with reliability.
Log in or sign up for Devpost to join the conversation.