Our project was inspired by the CSCI0220 staff theme of controversial sandwich. We took on the burden of answering this all-important question.
What it does
Through our app, we are able to classify any food item as a salad, toast, sandwich, taco, sushi, bread bowl, or calzone using deep learning.
How I built it
We created a CNN using keras to classify an input image as one of the 7 classes listed above. We trained our model using Spell's GPUs.
The frontend of the app is built with react-native and the backend is built with flask.
Challenges I ran into
Communication between frontend and backend sending images, very messy unlabeled image data sets.
Accomplishments that I'm proud of
Successfully deployed working CNN model on Spell's GPU. Successfully connected between react-native frontend and flask backend
What I learned
We learned how to integrate frontend and backend and how to deploy on Spell's GPU.
What's next for Cube Rule