🍴 Food Visualizer 🍴

This Project helps in classifying food items and obtain nutritional insights , restaurant and recipe recommendations.All of this is facilitated through the react native application which works on both android and ios

Table of Contents

• Introduction
• Demo
• GettingStarted
  • Prerequisites
  • deploying
• TechStack
• Visualizations
• Summary
• FutureWork
• Acknowledgments

Introduction

The user can interface through the app where he/she can take a picture/Choose an image from the gallery and then a prediction for the food is made, once that is done, we can get nutritional insights, restaurant and recipe recommendations on the dashboard.

We can see the architecture diagram of the entire system below:

Demo

You can download the app by clicking this link

GettingStarted

Prerequisites

install node package manager

pip install npm

install expo-cli

npm install -g expo-cli

install firebase

npm install -S firebase

Deploying

We use flask and app engine to deploy the website

setup google app engine

install gcloud from this link

After installing gcloud we need to setup app engine

gcloud components install app-engine-python

setup Third party libraries

we run this command inside the dashboard folder

pip install -t lib -r requirements.txt

TechStack

• App Development
  • expo (React Native)
  • Google Vision API
  • Google Firebase
• Dashboard
  • Google App Engine
  • Google Cassandra Cluster
  • Flask
  • Jinja2
  • Google Maps API
  • Chart.js
  • Bootstrap

Visualizations

We get insights for nutrients and recommendations for recipe and restaurants, We can see sample output for pizza below

Nutrients insights

insights for nutrients

insights on calorie values

insights on food fat

Restaurant Recommendation

Recipe Recommendation

Summary

So as we can see, given an image of any food item, we are able to classify the food item. Once the item is classified, we make 3 GET requests to our dashboard, where we pass the food label name, the dashboard receives this food label and does the following 3 tasks

• Query Cassandra with the food label to get nutrients information
• Query Cassandra with the food label to get top rated restaurants information
• Scrape recipe data for the identified food label

Once we receive data from cassandra to our dashboard, we then pass these values to our visualizations through jinja which is a template engine We can then view all our visualizations in one place.

FutureWork

We are planning to add support for user profiles in order to store previous data. We can see the previously scanned images and also the insights related to that. That way a user can lead a better life by making healthy life choices 😃

Acknowledgements

• react-native app