Ingredient lists on food packaging are often confusing and provide little meaningful information to consumers. Also, they are hard to read and contain many different chemical names unknown to the general population. We wanted to create an app to educate and inform individuals about the food that they are ingesting!
What it does
Takes a picture of the ingredient list of a food item, uses Firebase ML Kit Natural Language Recognition to recognize the text, uses the recognized text to search Wikipedia and extract relevant paragraphs containing keywords (e.g. safe, health), and displays that information in-app with an overall healthiness rating for the food item.
How we built it
- Used Node.js for backend development hosted on cloud functions and analysis of text
- Used Android SDK for app development
- Firebase ML Kit used for recognition of text and text extraction
Challenges we ran into
- Syncing backend with the android app
- Processing the text scan (able to convert to plain text but had a hard time parsing the list of ingredients)
Accomplishments that we're proud of
- Successfully using Android SDK to create an application that works!
What we learned
- Working with Android Studio
What's next for utrition
- Perhaps incorporating AR
- Expanding to search and filter in other websites besides Wikipedia
- Provide/Publish nutrition datasets