YourHome

Inspiration

We wanted to re-imagine what student living could look like and create an app that inspires better, cleaner, and more comfortable layouts. YourHome enables users to create a personalized bedroom based on their preferences, re-imagining bedrooms designing. This app would be very useful in expanding the user's choices for bedroom layouts by supplementing the selection process with the creativity of GAI.

What it does

YourHome is a home and bedroom design app where users can tinker with layouts, explore ideas, and get inspiration for transforming small or cluttered spaces into welcoming ones. By swiping on different layouts of bedrooms, the user can express their preferences for different designs and styles in bedrooms, all of which is put together by YourHome into a final personalized design recommendation generated through stable diffusion.

How we built it

We built the iOS app in Swift using concurrency and functional programming concepts, set up a Flask backend for requests for authentication and recommending images, and used CLIP models to embed images and create our initial database of bedrooms. The processed image into vector embeddings are stored in a SQL database, in which preferred bedroom embeddings are stored for personalized recommendations identified through cosine similarity and k-means clustering. When the user requests a personalized room, a vector representation of the past user preferences is processed through stable diffusion to generate an image of their personalized room.

Challenges we ran into

Connecting the backend to the frontend required multiple iterations and testing overnight. By testing different posts and get requests, we debugged all of the conflicts within the frontend and backend. Additionally, finding the vector representation based on past user preferences required creative workarounds.

Accomplishments that we're proud of

We built a fully functional iOS app within the hackathon timeline, delivered a working swipe system to browse design inspirations, and integrate backend + AI image embeddings into a single flow.

What we learned

We learned how to develop and deploy in Swift for iOS, how to connect a Python/Flask backend with a mobile frontend, and combining UI/UX design with AI-driven features. This project also accentuated the importance of collaboration. Without careful communication between the backend and the frontend teams to debug their respective Swift and Python codes, for example, the user interface would not able to obtain necessary information such as the user login information and the images of the bedrooms.

What's next for YourHome

We plan to Improve the UI/UX to make it more polished and user-friendly, expand our dataset to provide richer design inspirations, and implement deeper AI integration for personalized recommendations.