Inspiration
Started with a simple idea, what if we could use image recognition to simply find out something is keto-friendly or not?
What it does
Take a photo of your food with your phone, then send it to our message processing server, It will try then use image recognition software to determine the name of the food. It will then give you an estimate of the carbohydrate count to determine if it's keto friendly or not.
How we built it
We used Clarifai, an image recognition API and USDA's food database API. We used Flask a Python framework for our application and Twilio, a cloud-based messaging platform.
Challenges we ran into
Parsing through the black hole of a JSON file from the USDA's food database, trying to read in the target value. Making Clarifai's API and USDA's API work in unison in our Flask application. Narrowing the possibilities of the image recognition software to a single model, rather than a general one.
Accomplishments that we're proud of
Learning Flask and Python, learning how to integrate APIs in our application, using our phones as tools to maximize the potential of our application, teamwork, and above all, staying alive.
What we learned
How to work better as a team, planning out our development workflow, learning a whole new framework and language, Flask and Python, learning how to use APIs and Twilio.
What's next for Ketogram
Integration with a web application where users can make profiles and store their picture history online, allowing them to create meal plans based on the food they have eaten often.
Just text (956) 403-6287 and take a photo of your food!
Log in or sign up for Devpost to join the conversation.