Building a 2-in-1 intelligent tool to recognise ingredients and recommend recipes

Pantry-to-Plate buddy.ai

Problem statement:

Given the current food waste situation in Singapore (detailed in repo's ppt deck) and Singaporeans' time-starved lifestyle, how might we help households in Singapore to deal with excess groceries at home and reduce food waste?

Solution aim:

This project aims to supplement other food reduction efforts by providing an easy tool on utilising existing ingredients at home to whip up tasty meals. To this end, there will be two parts in this project: 1) an ingredient detection model, and 2) a recipe recommender system.

In this project series, there are 4 notebooks:

1. Data collection - recipe dataset (from http://www.themeatmen.sg)
2. EDA and recipe recommender
3. Collecting images of ingredients (from Fairprice, Roboflow, Flickr)
4. Custom training of object detection model (note: this is run on Google Colab)

Try the working demo on https://githubberst.pythonanywhere.com