Inspiration
1.4 billion tons of food are wasted every year
What it does
Identifies food in a refrigerator that is about to expire and provides recipes they can be used in so they don't get wasted.
How we built it
Used 4000 images of fresh and stale apples (provided by Kaggle) to train a feature recognition neural network. React.js was used to develop a user interface for mobile devices.
Challenges we ran into
- Understanding how machine learning works
- Debugging with React and javascript
- Finding suitable hardware for our idea
Accomplishments that we're proud of
- Trained a highly accurate neural network
- Created a beautiful user interface
- Gave birth to a product that could change the world
What we learned
- Translating ideas into real life is difficult
- Designing a product is more than just coding
What's next for Spoiler Alert
We will collect data about the entire decay process of apples to test the feasibility of predicting shelf life exclusively through computer vision, bypassing other physicochemical analysis. We will also experiment with different camera sensors to find one that is easily scalable and installable. We also plan to expand our product to other stakeholders within the food system (distributors, growers, restaurant owners).
Log in or sign up for Devpost to join the conversation.