We were inspired by the problems that we faced as first year students moving from different regions of the world.
What it does
Our app makes healthy food EasyToMakeO. Gone are the days of looking at your fridge and having no idea what to make. EasyToMakeO instantly scans your fridge to see exactly what ingredients you have available, and then scours through half a million recipes to find the best recipes for you.
How we built it
We first used Google’s teachable machine to train a quick Convolutional Neural Network on 30 different food items for a sanity check and then proceeded to fine tune the model and add multiple bounding box detections. Once we were happy with consistent detections of all items in our fridge, we moved on to the recipe hunting algorithm. We pick recipes that 1) use the most of available ingredients 2) require to buy the least number of new ingredients We then made a prototype of the app in Figma to model an optimized and easy to use interface
Challenges we ran into
The biggest challenge we faced was cleaning the dataset which contained half a million recipes and creating an algorithm to be able to search for specific keywords within such a dataset.
Accomplishments that we're proud of
We're excessively proud of being able to reduce computation time on the search algorithm and deploy it alongside an entire object recognition software trained with over 30 unique food items.
What we learned
What's next for Easy ToMako
In the future, we would cooperate with food sustainability and accessibility apps to connect students with the closest affordable options and promote purchase of groceries, as well as incorporate a delivery feature to boost convenience. We would also develop an algorithm to customize recipe suggestions based on your ratings and have the recipe suggestions be not only Easy ToMako, but also ideally tasty to eat!