Smart Closet

Home Screen
Closet Page

Inspiration

Our inspiration comes from a commitment to sustainable fashion. In recent times, fast fashion and unnecessary overconsumption of clothing is on the rise. Many times, consumers aren't able to be creative with the clothes they already have, leading to them buying clothes they don't need. Smart Closet aims to reduce the need for new clothes, inspiring creativity in existing clothes.

What it does

The application has multiple functions depending on the user's style needs. First, it allows users to upload their clothes to the application where the app will recognize the type, color, and fit of the clothing to recommend outfit combinations. Second, it allows users to swipe on outfit combinations, swiping right to save and left to reject. With this, the application also learns the users tastes to better recommendations.

How we built it

There are multiple components to our application. Firstly, our machine learning model that is able to detect types of clothing was developed using a CNN model using TensorFlow. This model was deployed using Terraform on a Flask Server. The UI was built on TypeScript and React Native, allowing for an intuitive user interface.

Challenges we ran into

We had a hard time connecting our services. Our API requests were not able to be connected to the front-end at all, and so we had to deploy it all on a web server and link it to our front-end that way.

Accomplishments that we're proud of

Making the UI look nice and smooth.

What we learned

To version control better, and make sure everyone is familiar with the tech stack.

What's next for Smart Closet

We aim to build more features into our Explore page. Some ideas include outfit recommendations on a budget or outfit recommendations for fitness. We also hope to eventually deploy to the App Store.