Room Shuffle
Generating furnishing solutions in real time.Room Shuffle is a machine learning assistant that helps you furnish your space, shows unexplored interior design possibilities and unlocks potential in the ways we use our everyday spaces.
What we did
- Prototype development
- Machine learning data training
- 3D environment design
A technical prototype for smart floor plans
Across a diverse set of themes, featuring technologies ranging from machine learning to spatial audio, SPACE10’s Everyday Experiments has explored ways to enhance our interactions with space and improve our everyday lives. We were invited to create four prototypes for the first Everyday Experiments series, one of which we named Room Shuffle.
Room Shuffle is a technical prototype that is designed not only to help you furnish your space, but to help you generate new ideas as well. It creates furnishing solutions based on common interior design rules; how much space furniture needs in relation to walls or other pieces, how furniture tends to be positioned in a room, and how different pieces are usually combined (how a sofa tends to sit beside a coffee table, or a dining room comes with chairs, for example).
Feeding Room Shuffle with predefined and stylistically coherent room-sets of furniture, we fit a series of probabilistic models for the occurrence, placement, rotation and padding of furniture in different room types.
Preview AI-generated suggestions for your space in 3D
Unlike data-hungry methods like deep learning, the probabilistic modelling of Room Shuffle can be fed as little as ten training examples. It is also a very flexible framework for learning, and can easily be adapted to more constraints in the future: modelling more levels of interior designs, such as wall or ceiling objects or general accessories, can easily be added down the line.
When using Room Shuffle, you can start designing in an empty room or take existing doors, windows, lights or furniture pieces into account. Generated furnishings can be viewed live in a three-dimensional (3D) representation of your space on your phone or tablet, allowing you to instantly witness how the suggestions would work in your physical space. A simple yet effective way of dreaming about your future interior.
Visualising clustering thresholds for grouping chairs into a "motif".
Tech Stack
- Python with Pydantic/mypy
- NumPy, SciPy and scikit-learn
- FastAPI