MealMate

Fridge that lacks, got our backs! We're here to help you with your meals. Just put in an image to our application, edit the food ingredients, and we'll make you a recipe with detailed instructions!

Comment

Inspiration

The inspiration for our project came from struggles to find or create recipes with the random items that we had inside our fridge. Therefore, we set out to create a platform that would allow users to submit images and receive recipe suggestions and nutrients

What it does?

MealMate takes in an image and classifies the food within that image. Then, it finds a recipe that uses those ingredients to make a delicious meal!

How we built it

We use two machine learning models, namely YOLOv8 (an open-source computer vision model we are using for Food Item Recognition) and Gemini (google's best large language model by far which we are using for generating a Recipe out of those food items according to the user's preferences). We built

Challenges we ran into

  1. Deciding on the API to use for OCR since the model required high fine-tuning and high epoch.
  2. Most of the APIs such as Cloud Vision API required a billing account which narrowed down the APIs we could choose from.

Accomplishments that we're proud of

Within the timeframe, our team managed to finish training both machine learning models to detect food and create recipes given a list of ingredients. Although we ran out of time to connect the model to the front end, we are still proud of learning and getting very close! For the backend, we used JavaScript and Flask to manage and connect the components to the Python file by creating a Post and Get API call.

What we learned

  1. The research in health and nutrition using the datasets for the daily diets of an average American helped us understand common food patterns.
  2. The project further improved our understanding of deep learning, OCR, and NLP.
  3. Along with that we got a chance to get some hands-on experience in Web Application development.
  4. Learned a lot about Hugging Face models that

What's next for MealMate

Add an eating-in or eating-out option in case you are too lazy to cook, but don't want to break the fitness streak. We suggest the nearest restaurants based on your cravings and fitness goals. Suggesting super foods in case you are feeling sick, to help you realize the power a meal carries to immunize your body and provide strength. Introducing you to lifestyle changes to make your fitness goals easier to reach. This contains but is not limited to introducing new diet plans, new dishes to incorporate necessary macronutrients etc.

Share this project:

Updates