Yummalyzer - The Food Analyzer

What the bite?! Yummalyzer is a Flutter based Android application designed to analyze food images and provide detailed nutritional insights, helping users see their food nutrition using Gemini API.

Comment

Inspiration

Everyone wants to eat right, but not everyone knows how to check. We wanted to create a simple way for people to analyze the food they eat just by taking a picture. Using AI and image recognition, Yummalyzer allows users to instantly get nutritional info from their meals.

What it does

Yummalyzer allows users to take or upload a picture of their food. It then uses AI to analyze the image and presents a nutritional overview and information such as how healthy the food is, any possible allergens the food may contain, how processed the item is, and any high sugar/sodium counts. The user can save their scans for future reference.

How we built it

We built Yummalyzer using Flutter/Dart for Android development. For image recognition, we used Google's Gemini API, to get structured nutritional data using carefully crafted prompts.

Challenges we ran into

  • Initial Setup: Getting the platform ready was tedious. First we had to pivot from Android Studio to Flutter after which we had a lot of trouble with dependencies and bad configuration.
  • Image Recognition: Figuring out how to send the image to the Gemini api was a challenge.
  • Data Formatting: Getting properly formatted response from the Gemini model required careful fine-tuning of the prompts in order to avoid causing potential errors from the received info not being formatted correctly.
  • Tracker: We had trouble with displaying all the scans that have been saved by everyone. We figured out how to properly display JSON information in a more human readable format.
  • Getting the Notes to display: Gemini was acting like a finicky black box at times. We had to do much tinkering with the AI get it to respond well.

Accomplishments that we're proud of

All of us are really proud to be able to present a well-polished application, using a language that a majority of us were unfamiliar with. On the personal side, one of our members was proud of being able to get JSON data formatted correctly on our tracker. All of us were very proud at getting the notes to appear more consistently.

What we learned

We learned about many aspects of flutter and the app development process, we learned how to send images gathered through camera hardware and send it to Gemini. We also learned how to better troubleshoot problems in general and to think more like a software engineer.

What's next for Yummalyzer - The Food Analyzer

A possible roadmap for Yummalyzer includes dietary suggestions based on user goals, such as weight loss or muscle gain. A personalized dashboard for users to be able to track long term dietary goals and user accounts for online syncing.

Built With

Share this project:

Updates