Nutrivision

The aim of this project is to create a comprehensive image classification and advice generation system which takes the images of food product and estimate the calories in that and then give appropriate amount one should take to maintain his/her diet.

Link to the frontend: Nutrivision-Frontend

Run Locally

Clone the project

  git clone https://link-to-project

Go to the project directory

  cd Nutrivision-Backend

Create and activate environment

  conda create -n nutrivision python=3.10
  conda activate nutrivision

Install dependencies

  cd environment
  conda install --yes --file requirements1.yml
  conda install --yes --file requirements2.yml

Start the server

  python3 detect.py
  python3 main.py
  python3 retrival.py

Environment Variables

To run this project, you will need to update the following variables in the retrival.py

GPT_API_KEY

GEMINI_API_KEY

ANTHROPIC_API_KEY

API Reference

Get all items

  GET :8001/nutritional_info

Parameter	Type	Description
dish	string	Required. The dish for which we want the nutritional values to be extracted.

Get item

  GET :8002/info_gemini/invoke

Parameter	Type	Description
'input':{'dish':dish}	json	Required. Internal call to fetch the data from gemini

  GET :8002/info_gpt/invoke

Parameter	Type	Description
'input':{'dish':dish}	json	Required. Internal call to fetch the data from gpt

  GET :8002/info_claude/invoke

Parameter	Type	Description
'input':{'dish':dish}	json	Required. Internal call to fetch the data from claude

  GET :8003/detect

Parameter	Type	Description
file	file	Required. Image of the food item for the classification

Screenshots

Demo

Nutrivision.mp4