HouseDiffusion API

HouseDiffusion: Vector Floorplan Generation via a Diffusion Model with Discrete and Continuous Denoising

Installation

1. Clone repo and install the requirements:

My implementation is based on the public implementation of HouseDiffusion.

git clone https://github.com/sakmalh/house_diffusion
cd house_diffusion
pip install -r requirements.txt

2. Initiate the Backend

uvicorn app:app --host 0.0.0.0 

• The pretrained model will be downloaded automatically from the Google Drive. pretrained

Endpoint

1. Request and Response Types

• To test the endpoint you could simply access http://0.0.0.0:8080/generate

{
  "nodes": [
    {
      "id": "string", # Nodes unique id   
      "room_type": "string", # Room Type in ['Living Room', 'Kitchen', 'Bedroom', 'Bathroom', 'Balcony', 'Entrance', 'Dining Room', 'Study Room', 'Storage', 'Front Door', 'Unknown', 'Interior Door']
      "corners": "string" # Number of Corners
    }
  ],
  "edges": [
    {
      "id": "string", # Edge Unique id
      "source": "string", # First Nodes id
      "target": "string"  # Second Nodes id
    }
  ],
  "metrics": true # If True provides the length and width of rooms in pixels
}

Deployment

• DockerFile is provided for hosting.

docker build -t {docker-repo}/backend .

Citation

@article{shabani2022housediffusion,
  title={HouseDiffusion: Vector Floorplan Generation via a Diffusion Model with Discrete and Continuous Denoising},
  author={Shabani, Mohammad Amin and Hosseini, Sepidehsadat and Furukawa, Yasutaka},
  journal={arXiv preprint arXiv:2211.13287},
  year={2022}
}