Inspiration
Over 15 million Americans struggle with food allergies. For these individuals, each meal can literally be a life-or-death situation. Our goal was to develop software that can help these people know what's in their food quickly and effortlessly, granting them a measure of assurance with the next bite.
What it does
Edible is a mobile app for Android. Users take a picture of dish, which then gets uploaded to our servers where it gets analyzed. A few seconds later, the app displays what ingredients likely went into the making of that dish, and whether any are food-allergens. The app can also recognize meat and animal products for vegetarians and vegans, too.
How we built it
The front-end was developed using Android Studio. The backend is written in Python and uploaded to an App Engine from Google Cloud Platform, where it carries out server-side computation and communicates results with the app. To recognize ingredients, we used a machine learning project from MIT called im2recipe, focused on deducing the recipe given an image of food.
Challenges we ran into
The biggest challenge was figuring out how to get the Python code to run properly on the App Engine instance. After setting up a Cloud Endpoints API, we communicated successfully with our server, but regretablly never got the Python code itself to run properly on the server. The code works fine locally, but for some reason fails when deployed to App Engine.
Accomplishments that we're proud of
We're excited that we've got an approximately fully-functional front-end and back-end (we've even got the logo), the only piece missing is connection between them. Alas, if only we could get the App Engine to function correctly, but we're proud nonetheless that everything else works okay :)
What we learned
We learned plenty about the Google Cloud Platform, the im2recipe project, and Android developement. It's been an awesome 20 hours of intense googling!
What's next for Edible
Next is getting the App Engine to work properly. Once that's in place, the app would become more or less a functional prototype. After that, there's definitely a need to do some serious code-cleanup and polishing. Future features like refining the ingredients list and improving accuracy are also a must.
Log in or sign up for Devpost to join the conversation.