Frigotopia

Landing
Fridge
Fridge Options
Manually Insert Items
Adding item
Receipt
Recept Validate
Image Validate - Ripeness
View Item - Update
Recipes
View Recipes
Notification - Time to use food

Inspiration

You know the feeling of having those random days where you look through your entire fridge and find food that has gone bad that you have completely forgotten about and ended up having to throw. Us too, actually I can bet most people can relate to that feeling, and certainly, our team could. Sadly food wastage is a big problem in North America with over 40% of all foods going to waste with the majority of that coming from households. It was for this reason our team decided we need a solution to this in order to reduce food waste in households.

What it does

At the core, our app simply tracks the items that are within your fridge. This can be used to help consumers be more aware of the contents of their fridge. Either it be while shopping to ensure they don't purchase more than they need or to use it to get a better view of their fridge contents. On top of that, our app also allows for more information about your different food items to be tracked, such as date of purchase, days until expired, quantity, and more. With the information, our app will send you notifications to remind you when you have something that will go bad soon and will even suggest to you recipes you can use to use up the soon to go bad food!

In order to track what is in your fridge, our apps provide 3 methods to add food items.

Manual Add: Select from a list of different food products
Receipt reader: Read the contents of your receipt from your purchase using google-vision
Food scanner: Scan your groceries and let our AI determine what you bought and its ripeness

How we built it

Our application is a mobile app that was built using flutter in order to ensure that it will be compatible with both Android and IOS. The backend of our application leverages Flask for our API in order to ensure smooth communication with the front-end.

For the food detection, we are leveraging OpenCV and YOLO in order to determine the food present within the image. Then we use (for now only with Bananas) a custom built CNN model that determines the ripeness of the fruit using Tenserflow and Keras. For the receipt reading, we are leveraging google vision in order to extract the text from the image.

Challenges we ran into

The main challenges our team faces were in the fact that we are all new to the many technologies that were used for our app. None of us have ever used flutter before so the setup and the entire front-end process was confusing for us all. On top of that, most of the technologies we used on the backend as well were new technologies, especially YOLO object detection which took lots of time to properly understand how it functions.

Accomplishments that we're proud of

We are really proud of the app as a whole. The app is not close to polished, but we believe it to be very useful in many situations. We really hope that we can finish this app as we believe it will be the first steps in reducing food wastage in North America.

What we learned

During the hackathon, we have learned a lot about the different technologies we were using as mentioned before. On top of that, we all learnt how to better manage to work in a team under lots of time pressure as well as under a lot of stress due to a lack of sleep.

What's next for Fridgotopia

Our team would love to continue polishing the app to get all of its base features working flawlessly. Then we believe the app would have lots of value as a free to download application that can have impact in all houses accross North America.