NutriVision

upload photo page
calories statistics page
meal suggestion page

Inspiration

We were tired of manually entering details about the food we're eating. Remember the initial excitement of tracking your food intake? It quickly fades, replaced by the tedious task of logging every bite. We've all been there. But what if tracking your nutrition could be fun?

Introducing a revolutionary app that turns calorie tracking fun and easy. Simply snap a picture of your meal, and our intelligent technology analyzes the image, revealing the calories and nutritional breakdown in seconds. No more endless typing!

But we didn't stop there. We understand the fridge-staring dilemma - those random ingredients begging to be transformed. That's why our app also serves as your personal meal prep hero. Based on the ingredients you've snapped, we'll generate creative and delicious recipe suggestions, helping you get the most out of your groceries and avoid food waste.

What it does

Streamlined Nutrition Tracking:

Users capture and upload food images, eliminating manual ingredient entry.
Our intelligent object detection system, identifies items with high accuracy.
Users can confirm or edit findings for guaranteed precision.

API Integration for Rich Data:

Confirmed ingredients are fed into a robust nutrition API, unlocking comprehensive nutritional data.
Calories, and measures like sugar and fat are readily accessible within the app.
Data is stored in MongoDB for further retrieval and active learning

User-Friendly Nutritional Insights:

The application presents a clear and concise breakdown of nutritional components.
Uses graph to track progress

AI-Powered Meal Prep Assistant:

Gemini LLM analyzes existing ingredients, suggesting innovative recipe ideas.
This functionality minimizes food waste and fosters culinary exploration.

How we built it

Food Detection and description

Leveraged object detection techniques to accurately identify food items uploaded by the user
Used Gemini to generate apt descriptions to feed the API

Generating Nutrional insights using Nutrition API

Used nutrition API to generate detailed nutritional breakdowns and calorie counts.

Database management

Implemented a robust database management system using MongoDB for efficient data storage.
Stores user-uploaded images, food descriptions, and nutritional values retrieved from the API.
Integrates seamlessly with meal prep suggestions for a well-rounded approach to healthy eating.

AI Recommondations

Harnessed LLM to generate meal prep suggestions
Focused on optimising ingredients and proposing healthy options

Front- End

Modern tech stack featuring Next.js and React.js for a user-friendly and interactive experience.
Jotai for seamless state management.
Tailwind CSS and TypeScript for a visually appealing and scalable design.

Back-end

Utilized the well-established Flask framework for its flexibility and ease of use.
Integrated the Python Imaging Library (PIL) for effective image processing.

Challenges we ran into

Configuring mongoDB
Co-ordinating between backend servers and frontend to identify cause of bugs
Dealing with compatibility issues between libraries

Accomplishments that we're proud of

Getting backend and frontend to co-ordinate well
Leveraging AI for personal custom recommendations
Interactive and aesthetic UI/UX

What we learned

Storing and retrieving data hosted on MongoDB atlas and connecting it to back-end servers.
Using LLMs to generate accurate image descriptions and vice-versa using heavy prompt engineering

What's next for NutriVision

More statistical insights and personalized recommendations based on gut health
More personalised tracking and recommendations by integrating user health conditions and goals
running models without API for faster response times