One Bad Apple

Inspiration

One Bad Apple is inspired by one of the United Nations 17 Sustainable development goals, specifically: Good Health and Well-Being.

What it does

One Bad Apple is a website that allows users to take a picture of a fruit (either an apple, a banana, or an orange), and tells the user whether or not their fruit is safe to eat. This helps address those in impoverished communities, as often those people are forced to choose between eating potentially unsafe food or not eating, potentially leading to starvation.

How we built it

Front-end was created with Next.js with TypeScript and Tailwind CSS.
Back-end was created with Python (Flask)
Neural network was made with Keras using the TensorFlow framework
Machine learning model data pulled from Kaggle.

Challenges we ran into

The most persistent challenge that our group faced was in creating the model. Firstly, we did not anticipate how long it would take to train the model (for reference, it took a couple hours, meaning that we lost time that could have been used to testing). In addition, communicating the results from the model to the frontend posed a challenge as our group had minimal experience with Flask.

Accomplishments that we're proud of

Something we are most proud of, is the UI that we created, as the user is able to upload an image of their choice, and Rot or Not displays this image, and tells whether or not their food is safe to eat.

What we learned

We severely underestimated the amount of time and resources it would take to train our own model. Thus, we learned that we should have used another resource, such as an existing method to train a model, instead of having to train our own (sacrificing so much time).