-
-
The start page that displays a camera button, and a dark submit button when the image is ready for analyzing.
-
The results page after a successful analysis.
Inspiration
The inspiration for CampNav came from my [Jerry] experiences at other colleges. I am usually attempting to become familiar with the buildings throughout the campus. Sometimes, the names of the buildings are not revealed from a certain point or angle. I thought, "What if I can build an app that will identify the buildings for others who are new to a college campus?"
What it does do
The app allows a user to take a picture of a campus building in Stony Brook University. The picture is then sent off to a server to be processed by a machine learning algorithm. A result including the name of the building and its score is returned and displayed in a modal.
How we built it
The team was split into two groups, the backend developers and the frontend developer. The backend developers built a Flask API that uses Google's AutoML tool to receive an image, process the image and return a suitable result. The frontend developer built a React application for the user interface.
Challenges we ran into
We ran into several challenges. The first challenge was getting a suitable API for the image classification. We were originally going to use Clarifai; however, due to limitations it proved difficult. Instead, we used AutoML by Google which proved to be a more seamless process for our use case. The second challenge we ran into was building an optimal interface between the client and the server. We had to utilize AJAX technologies to be able to build a suitable interface for both sides of the application. The third challenge we ran into was having the app work on mobile. Mobile devices is our target for the application, and so we placed some importance on the app's functionality on mobile devices.
Accomplishments that we're proud of
We are proud of having a working image classification model that works with a few buildings on Stony Brook's campus.
What I learned
Kelvin - "I learned how to make a simple Flask application." Jeff - "I never built anything meaningful with Python, so it was a great learning experience." Jerry - "AJAX!" Spencer - "I learned Flask."
What's next for CampNav
We could continue adding images to the our AutoML model. We can also introduce accounts where a history can be recorded of the images and results for the user. We may also experiment with AR so that the user's camera can hover a label on a building with the appropriate results.
Log in or sign up for Devpost to join the conversation.